This is the Final Progress Report for WP2 of W3C-LA. The work is described in relation to the Tasks defined in the original Technical Annex together with modifications agreed at the May 1998 Review.
The management of the Project has been integrated with the management of the World Wide Web Consortium (W3C). W3C is hosted by three organizations:
Each host is in charge of the activities in its corresponding continent. INRIA is also responsible for Africa. During the course of the W3C-LA, W3C has introduced a new level of operations called Offices. Offices are responsible for a better communication channel between the consortium and their respective countries. Offices undertake Marketing programmes to raise new membership. They are also the primary point of contact for the existing members. Offices are vendor-neutral organizations, W3C members and should have a good network of relationships within their country.
Jean-François Abramatic is the Chairman of W3C and has been the Project leader for W3C-LA for INRIA. Bob Hopgood, Head of the UK Office at RAL has been the Project leader for RAL.
Management of the project has benefited from the contribution of W3C management people : Vincent Quint, User Interface Domain leader and Deputy Director for Europe, Philipp Hoschka, Architecture Domain leader, Josef Dietl, Head of Member Relations, Daniel Dardailler, Technical Lead of the Web Accessibility Initiative and Sophia-Antipolis site manager.
Each workpackages were under the responsibility of a workpackage manager. Dissemination of W3C results was implemented by demonstrators developed on top of the W3C Open Source Software platform : Amaya, the client testbed, Jigsaw, the server testbed, Libwww, the library testbed.
At the end of the project, some demonstrators have reached a sufficient maturity to be made available to the general public. They are :
The Project has developed in conformance to the programme. It has been extended by three months which allowed to start the deployment of Offices in the Mediterranean area. Contacts and first plans have been established in Tunisia and Morocco.
| WP2 | Marketing and Dissemination | Audit of European companies, W3C Members and Web users completed as a guide to future W3C Members. W3C Offices established in the UK, Germany, the Netherlands Greece and Sweden. A sixth Office was established in Italy near the end of the Project. Initial set of literature defined and used at both Road Show Events and Symposia. The literature was updated regularly with one major update after a year. A set of Demonstrations for use at Road Shows and Symposia developed early on and extended as the project proceeded. Many Road Shows given in 9 countries (UK, Germany, France, Sweden, the Netherlands, Greece, Italy, Austria, Belgium). First Symposium took place ahead of schedule and Report written. Three more Symposia given in Sweden, Germany and the Netherlands. First Disseminator Advice Workshop delivered on time and repeated in other countries First Demonstrator Dissemination Event delivered on time. Further events held in Holland and Sweden. Many Awareness talks given throughout Europe. Significant increase in the number of European members, particularly SMEs. |
Due Date |
Available |
Resp. |
Type |
Content |
01.02.98 |
Yes |
RAL |
M2.1 |
First Awareness Symposium & Report |
01.05.98 |
Yes |
RAL |
M2.2 |
First Demonstrator Advice Workshop |
01.01.99 |
Yes |
RAL |
M2.3 |
First Demonstrator Dissemination Event & Report |
Technical Annex |
Q1 |
Q2 |
Q3 |
Q4 |
Q5 |
Q6 |
|||||||||||||
T2.1: Audit and Analysis |
|||||||||||||||||||
T2.2: Market Access/Delivery Plan |
|||||||||||||||||||
T2.3: Define Web Awareness Material |
|||||||||||||||||||
T2.4: Produce Brochures |
|||||||||||||||||||
T2.5: Produce CD-ROMs |
|||||||||||||||||||
T2.6: Specify Road Show |
|||||||||||||||||||
T2.7: Develop Demonstrator Shell for Road Show |
|||||||||||||||||||
T2.8: Define Road Show Material |
|||||||||||||||||||
T2.9: Develop Specific Road Show Demonstrations |
|||||||||||||||||||
T2.10: Run Road Shows |
|||||||||||||||||||
Information concerning major companies in Europe that would be interested in W3C were collected with the UK as the first in-depth study. An analysis of existing members of W3C was also done to analyse which market sectors were covered. The top R&D spending companies in Europe were analysed to identify potential new members. A detailed analysis of US web sites in specific applications areas was done to see which sectors were likely to be the major beneficiaries from W3C-LA in Europe. The UK was used as the initial country for most of the activities and that experience was used in focusing the other European activities. The initial conclusions were summarised in the internal deliverable ID2.2.
The initial focus was on defining a set of events where a W3C-LA stand could be located and on producing the necessary stand material. A mock-up of the stand was done at RAL and, after modification, the first production stand was put together in time for EITC '97. Trial stands had already been manned at Thessaloniki in Greece and in Berlin (OpenNet'97).
The three main events organised by W3C-LA WP2 were the Awareness Symposia early on, the Demonstrator Advice Workshops mid-term and the Demonstrator Dissemination Events later on. These are described later in this Report.
In addition, Office staff have given a number of talks aimed at raising awareness of W3C-LA and the benefits of joining W3C. These are a combination of presentations to individual companies and presentations at Conferences or Briefing Meetings. As an example, the talks given by RAL are listed in Annex C.
A Monthly Newsletter was sent out from January 1998 initially to the attendees at the UK Symposium. By making contacts with over 1000 UK companies, the mailing list for the Newsletter rose to about 800 organisations in the UK who positively asked to receive it and future information about W3C-LA. The Newsletter was made available to the other European Offices and late on in the Project it has been converted to a Web Version and made available throughout Europe via the local Offices.
At the May Review, it was pointed out that Marketing is good for large companies but in Europe you need to convince SMEs. The belief was that the approach already adopted would not scale sufficiently to hit all the relevant SMEs in Europe.. A mass awareness campaign was needed that works outside the IT area.
In order to do this, each Office has built up an email list of specialist newspapers and journals to which W3C Press Releases are sent on a regular basis and translated into the local language where it is thought to be necessary. The major press is handled through the W3C Hosts.
Following the suggestions at the May review meeting, RAL, for example, sent out mail-shots to over 2000 SMEs (involved with the internet and web market) and to over 100 UK Press contacts. The Press contacts were formalized into an electronic mailing list to which W3C information is sent including W3C Press Releases.
The material required was broken down into:
Three initial brochures were produced:
The Who is W3C Brochure was updated twice through the life of the Project while the other two stayed relevant throughout. About 40 flyers were produced by October 1997 and these were the basis of material available at the early Road Show Events and responses to queries. Annex D lists the ones developed. A minor update of these was made about 9 months into the project and a complete rewrite around December 1998.. The text is stored centrally on a server at INRIA and all W3C Offices can download material and print it as required.
A special 2-page flyer in German was produced for CeBIT 97.
A Primer for CSS1 was produced early on.
The material produced early on is available in the deliverables (internal) ID2.4 and ID 2.5.
During the project Validators for both HTML and CSS have been made available on the W3C Web Site. There is also a comprehensive set of Accessibility Guidelines. An RDF Presentation system has been developed.
An Alpha testing of a CD-ROM was produced for the demonstrations used at EITC 97. A Beta test was done for the round of Symposia in March/April 1998. The first CD-ROM as a deliverable was distributed widely in Germany during the early part of 1998. It did not prove to be the best way of disseminating W3C code partly due to the explosion in web sites particularly by SMEs. In consequence, most organisations with an interest in the Web had the ability to download direct from the W3C web site. In consequence, downloading has been the most appropriate delivery vehicle for the WorkPackage deliverables.
A portable stand was purchased at the start of the Project. This can be carried by a single person on a plane so it was feasible for individuals to give both presentations and mount a stand at a remote site.. The panels for the four main boards were designed and used initially at EITC '97 and the London Symposium. The information was reviewed in the light of the experience at these two events and a second set of panels were used for the next three Symposia in March/April 1998 and subsequent Road Show events. A copy of all the material was kept in each Office for local use. The Road Show Stand consists of four main panels covering:
Annex A lists all the events where Road Show events have taken place during the project.
A set of guidelines were established with the aim that each demonstrator should adhere to these guidelines. In practice, it was found that demonstrations were under constant development and, early on, were dependent on specific browsers or machine configurations. In consequence, the concept of a shell was abandoned in favour of demonstrations that could be mounted on PCs and were self contained. It was then possible to have several demonstrations on the same system. The standard slide show software developed by W3C was used in some cases to produce a harness for the demonstrations.
Coupled with the stand material, each demonstration has a standard set of four Panels with appropriate paper handouts. One panel gives the basic information. A second is a schematic showing the demonstration. The third gives the technical details. The fourth indicates the problems solved by the demonstration.
The material used for the Road Show, the Symposia and general awareness were much the same. How the various pieces were used was different. The emphasis on 2-sided flyers, which could be mixed and matched in folders, allowed the material to be used in a variety of ways.
A meeting was held at Sophia Antipolis on 6 November, 1997 to define the initial set of demonstrations to be used at EITC '97. A second meeting at the beginning of February refined these for the next round of Symposia. A third review took place before the Demonstrator Dissemination events near the end of 1998. The demonstrations gradually used more of the results from the W3C-LA workpackages as the Project progressed. The set of demonstrations were:
Funding was available to run about 10 events in Europe. The agreed split was to be 4 Symposia, 4 intermediate events and 2 Demonstrator Dissemination Events.
A schedule for Symposium events was established with the first taking place in London on 3 December, 1997. This was successful with well over 150 attendees. Based on this, further Symposia took place in Sweden, Germany and the Netherlands between 30 March, 1998 and 2 April, 1998. Annex B gives the number of attendees at each event. A different style of event took place in Greece in September 1998. This was more tutorial in nature.
Technical versions of the Symposia were given in several countries after the initial round. The first of these was held in the UK at RAL and repeated elsewhere in Europe.
4 September 1998 - SMIL Tutorial and Hands-on Workshop at RAL.
The first Demonstrator Dissemination Event took place at RSA on 8th December 1998.
This followed on from the 1997 W3C Awareness Event at the same venue, by switching the emphasis to new and forthcoming developments in Web technology, user experience of these developments and the commercial opportunities that these offered. This provided a more technical and business flavour to the event. The talks were backed up with a full range of demonstrations, by user organizations as well as from INRIA and RAL. The deliverable M2.3 giving more detail on the event is available. Reuters and RivCom gave good presentations showing how they are using the latest W3C Recommendations as essential tools in their business strategy opening up new areas for their products.
Subsequent events took place in the Netherlands and Sweden. The event in Stockholm focused on the interest in the use of the Web from Mobile Devices and was the most well-attended event during the life of W3C-LA. The attendance list is given in Annex E to illustrate the breadth of attendees. This particular area had not even started when W3C-LA was instigated in 1997.
Some talks from previous events were repeated at the Opening of the Italian Office in April 1999.
Below is a factual record of how the various Offices have done in recruiting new members. The UK Office opened earlier than the other Offices by 3 months so you would expect the results there to be better than elsewhere.
The two major reasons for increasing European participation are:
The Total reflects the first of these and The Weighted Total, which uses the ratio 10:1 between Full and Affiliates (which is the differentiation in terms of cash received) reflects the second.
To give a measure of the effectiveness of the Offices in recruiting members, the results from the Offices as a set compared with the countries in Europe that did not have an Office.
At the mid-term Review of W3C-LA the project was asked to concentrate on the SME sector and thus not aim for maximum revenue to W3C. In consequence, the Weighted Total is less important than the Total in terms of assessing performance.
The set of countries where Offices have been established are:
A 6th Office in Italy has only just been established.
The W3C European Host is in France. Other European countries with W3C Members are:
In the Tables below, Overall includes all European countries. Other is the set of countries that do not have an Office or a W3C Host.
These figures indicate that the effect of the Offices overall was significant in attracting new members. There was a 78% increase in SME Membership across Europe and a 175% increase in those countries with Offices.
Events have been run in countries other than those where Offices exist and many attendees at meetings and events were from countries without an Office. Thus the effect that the Offices have on neighbouring countries is difficult to assess.
The major problem has been the failure of French companies to renew their Membership. However, as the thrust was to make the European involvement more even. The number of Members in Europe was originally over 50% from France (weighted total) and this has now dropped to 33% so that the balance within Europe is now much better.
Members |
Aug97 |
Feb99 |
Left |
Joined |
% |
| Full | 30 |
28 |
7 |
5 |
-7% |
| Affiliate | 33 |
59 |
5 |
31 |
+78% |
| Total | 63 |
87 |
12 |
36 |
+38% |
| Weighted | 333 |
339 |
75 |
81 |
+2% |
Members |
Aug97 |
Feb99 |
Left |
Joined |
% |
| Full | 16 |
10 |
6 |
0 |
-62% |
| Affiliate | 12 |
11 |
4 |
3 |
-8% |
| Total | 28 |
21 |
10 |
3 |
-25% |
| Weighted | 172 |
111 |
64 |
3 |
-35% |
Members |
Aug97 |
Feb99 |
Left |
Joined |
% |
| Full | 3 |
4 |
1 |
2 |
+33% |
| Affiliate | 1 |
6 |
0 |
5 |
+500% |
| Total | 4 |
10 |
1 |
+150% |
|
| Weighted | 31 |
46 |
10 |
25 |
+48% |
Members |
Aug97 |
Feb99 |
Left |
Joined |
% |
| Full | 0 |
0 |
0 |
0 |
0% |
| Affiliate | 1 |
1 |
0 |
0 |
0% |
| Total | 1 |
1 |
0 |
0 |
0% |
| Weighted | 1 |
1 |
0 |
0 |
0% |
| Members | Aug97 | Feb99 | Left | Joined | % |
| Full | 3 |
3 |
0 |
0 |
0% |
| Affiliate | 2 |
5 |
0 |
3 |
+150% |
| Total | 5 |
8 |
0 |
3 |
+60% |
| Weighted | 32 |
35 |
0 |
3 |
+9% |
Members |
Aug97 |
Feb99 |
Left |
Joined |
% |
| Full | 1 |
1 |
0 |
0 |
0% |
| Affiliate | 2 |
3 |
0 |
1 |
+50% |
| Total | 3 |
4 |
0 |
1 |
+33% |
| Weighted | 12 |
13 |
0 |
1 |
+8% |
Members |
Aug97 |
Feb99 |
Left |
Joined |
% |
| Full | 2 |
2 |
0 |
0 |
+0% |
| Affiliate | 6 |
18 |
1 |
13 |
+200% |
| Total | 8 |
20 |
1 |
13 |
+150% |
| Weighted | 26 |
38 |
1 |
13 |
+46% |
Members |
Aug97 |
Feb99 |
Left |
Joined |
% |
| Full | 9 |
10 |
1 |
2 |
+11% |
| Affiliate | 12 |
33 |
0 |
22 |
+175% |
| Total | 21 |
43 |
1 |
24 |
+105% |
| Weighted | 102 |
133 |
10 |
42 |
+30% |
Members |
Aug97 |
Feb99 |
Left |
Joined |
% |
| Full | 5 |
8 |
0 |
3 |
+60% |
| Affiliate | 9 |
15 |
0 |
6 |
+66% |
| Total | 14 |
23 |
0 |
9 |
+64% |
| Weighted | 59 |
95 |
0 |
36 |
+61% |
European Equivalent Full Members (France, Offices, Other)
World-wide Equivalent Full Members
The W3C-LA contract has been extended to allow for initiating W3C presence in the Mediterranean area. Efforts related to this topic are described in an Appendix to the W3C-LA contract :
Appendix 1- Internationalization : Arabization of W3C Open Source Software
Appendix 2 - Dissemination : Opening Offices in Morocco and Tunisia
In the three-month period (February - April 1999) of the contract extension, a plan for the Mediterranean actions has been developed based on initial experiments and meetings.
The arabization of Amaya, the W3C Open Source Client, has been included in the Amaya workplan based on experiments. The support of the Arabic language is the first experiment we did with non ISO-Latin languages in Amaya. We have chosen Arabic because it allows to deal with some interesting issues and because we have Arabic speaking people in our team. Arabic allows to experiment a new writing direction (RTL), this involves new new formatting and hyphenation issues. It also allows to use a new set of characters. Furthermore, Arabic is an alphabetic language and it does not need a complex input method.
The workplan is described in the document entitled "Globalisation of Amaya, the W3C Web Client : the Unicode version"
URL : http://www.inrialpes.fr/opera/people/Ramzi.Guetari/Papers/international_amaya.html
Regarding the arabization, three main items have been identified :
Leveraging the efforts made in opening offices in Europe, the Appendix to the W3C-LA contract proposed to open Offices in the Mediterranean area.
A series of meetings were held with the goal to extend the successful operation of W3C Offices into other regions, notably the Arab world. Two W3C Offices will be established, in Morocco and Tunisia respectively, each operated by a Consortium of two institutions in Morocco and Tunisia.
In Morocco, the consortium associates
In Tunisia, the consortium associates
The four organizations will sign
1. The W3C Membership agreement (each partner is expected to become a regular member of W3C)
2. A renewable contract with INRIA about the promotion of W3C and new technologies ("Office Contracts")
In order to deploy and demonstrate Web technologies in the Arab world, suitable tools and content have to be produced first. To achieve this, three main tasks have been identified and should become the subject of a Subcontract :
The consortia in Tunisia and Morocco will equally contribute to tasks (1) and (2) above, and each Consortium will operate a mirror site.
Based on these achievements, promotional activities (including continuation of the mirror site) should be undertaken in the scope of the Office Contract. Notably an official opening and a workshop on arabization have been discussed, experiences from other W3C Offices (Newsletter, etc..) mentioned. As the Office Contract will mention the participation of the W3C Office in the membership fees generated by the Members in their country as revenue stream, additional sources of funding are encouraged within the limits of vendor neutrality.
In order to kick off the effort, the following cornerstones have been agreed pending confirmation from other affected parties :
The change in Membership of the three sectors over the course of W3C-LA is given below
Area |
Full Aug97 |
Full Feb98 |
Changed |
Affiliate Aug97 |
Affiliate Feb98 |
Changed |
Americas |
33 |
47 |
+42% |
104 |
142 |
+36% |
Europe |
30 |
28 |
-7% |
33 |
59 |
+79% |
Asia |
13 |
14 |
+7% |
12 |
27 |
+125% |
Total |
76 |
89 |
+17% |
151 |
228 |
+51% |
The loss of Full members in France clearly effects these figures. The Figure below shows that the expansion in the USA continues to increase more than linearly whereas the increase in Europe is much more modest. The major push made by France to increase membership has not been repeated by the more modest funds available in WP2 of W3C-LA.
The encouraging statistic is that the number of new Affiliate members in Europe is rising at more than twice the rate of new Affiliates in the Americas. If we just look at those countries where we have European Offices, the rate of increase in Membership is four times the growth in the Americas.
It should also be realised that any activity of this type has an in-built lag and the true value of W3C-LA will not be known until at least another year. The awareness of W3C in Europe overall has risen significantly and that in itself will be beneficial in the long run and should encourage more European participation by one route or another.
Growth in Affiliate Members World-wide
We believe that W3C-LA WP2 has achieved the following:
The major disappointments are:
We are confident that the Membership of W3C in Europe will continue to rise even though WP2 of W3C-LA completed several months ago.
(On average, 5 Demonstrations plus Stand at each Event)
Date |
Location |
Event |
1-5.10.97 |
Thessaloniki |
INFOSYS '97 |
24.11.97 |
Brussels |
EITC '97 |
10-12.11.97 |
Berlin |
OpenNet '97 |
20.01.98 |
Amsterdam |
WebCHI, CWI |
19-25.03.98 |
Hannover |
CeBIT '98 |
16.03.98 |
London |
SET '98 (with CCTA) |
16.06.98 |
Bradford |
BCS MultiMedia |
26-30.06.98 |
RAL |
Open Week |
15-17.09.98 |
Newcastle |
UK Webmasters Conference |
23.10.98 |
London |
DTI/IETT Event |
16-17.11.98 |
Leeds |
MINA |
28.11-02.12.99 |
Vienna |
IST '98 |
(Talks, Stand, and Demonstrations)
Date |
Location |
Attendees |
03.12.97 |
Royal Society of the Arts, London |
174 |
30.03.98 |
Electrum Centre, Stockholm |
100 |
01.04.98 |
GMD Schloss, Sankt Augustin, Bonn |
138 |
02.04.98 |
Conference Centre, Utrecht |
122 |
17.07.98 |
Technical Workshop, RAL |
86 |
08.12.98 |
Royal Society of the Arts, London |
163 |
24.03.99 |
Electrum Centre, Stockholm |
274 |
22.04.99 |
Pisa |
110 |
Date |
Place |
Activity |
Speaker |
07.08.97 |
RAL |
CCTA |
Bob Hopgood |
21.10.97 |
Farnborough |
Post Office |
Bob Hopgood |
09.11.97 |
London |
BBC |
Bob Hopgood |
11.11.97 |
London |
Peer Association |
Bob Hopgood |
12.11.97 |
Leeds |
Halifax |
Bob Hopgood |
30.01.98 |
Milton Keynes |
Abbey National |
Bob Hopgood |
16.03.98 |
London |
SET '98 |
Bob Hopgood |
20.03.98 |
London |
BCS Multi-Media SIG |
Michael Wilson |
27.03.98 |
Leeds |
EG UK '98 |
Bob Hopgood |
11.05.98 |
London |
GIS Internet Forum |
Bob Hopgood |
12.06.98 |
London |
BBC |
Bob Hopgood |
16.06.98 |
London |
Net R&D '98 |
Brian Matthews |
16.06.98 |
Bradford |
SMIL |
Michael Wilson |
18.06.98 |
London |
XML |
Brian Matthews |
19.06.98 |
London |
Poptel |
Bob Hopgood |
10.11.98 |
Solihul |
TICAD '98 |
Brian Matthews |
16.11.98 |
Leeds |
MINA |
Michael Wilson |
11.12.98 |
London |
Brunel University |
Bob Hopgood |
17.02.99 |
Birmingham |
CALS Industry Forum |
Damian MacRandal |
13.04.99 |
Bradford |
Digital Media Futures |
Bob Hopgood |
The initial set of 2-page flyers used at road show events and for responding to requests for information.
Architecture Domain |
European Members |
PNG |
CSS |
HTML 3.2 |
Privacy |
Content Design Demonstrator |
HTML 4.0 |
Real Time |
Hyperglossaries Demonstrator |
HTTP 11 |
Security |
Labelling Demonstrator |
HTTP NG |
Technology and Society Domain |
MultiLingual Demonstrator |
I18N |
User Interface Domain |
Performance Demonstrator |
IPR |
W3C-LA |
Synchronisation Demonstrator |
JEPI |
WAI |
VectorGraphics Demonstrator |
Jigsaw |
Who is W3C |
WebCollections Demonstrator |
Math |
Why Join |
Workflow Demonstrator |
Metadata |
XML |
DOM |
Mirror Sites |
XML, HTML and RDF |
DSig |
PICS |
Revised Set Produced in December 1998
The set used for the events near the end of 1998 and early 1999 were closely aligned with the Activity statements within W3C.
Amaya |
Mathematics |
The Web in 1998 |
Web Characterisation |
Metadata |
1998/99 Web Enhancements |
Document Object Model |
Mobility |
The Web of the Future |
Digital Signatures |
Privacy |
Joining W3C |
Electronic Commerce |
Security |
W3C Recommendations |
Graphics |
Style sheets |
W3C Members |
HTML |
Synchronised Multimedia |
Graphics Demonstrations |
HTTP |
TV and the Web |
Workflow Demonstrator |
HTTP-NG |
WAI-IPO |
SMIL Demonstration |
Internationalization |
WAI-Technology |
Hyperglossary Demonstrator |
Jigsaw |
XML |
Thomas Abrahamsson |
Nokia Svenska |
Thomas Carnehult |
Telia Prosoft |
Richard Granberg |
Pro Innovateam |
Joakim Adler |
Mmt |
Joakim Cederblad |
Crescendo Communications |
Patrik Granström |
Nordbanken |
Rosie Adler |
Mmt |
Didier Chincholle |
Ericsson Radio Systems |
Andreas Gref |
Enator Industriteknik |
Magnus Agnarsson |
Telia Infomedia Reklam |
Eva Dahlberg |
Reston |
Olof Gunnarsson |
Arjano Data |
Arnold Andersson |
Handelsbanken |
Bo Dahlbom |
Informatik Viktoria |
Maria Gure |
Appius |
Johan Andersson |
Mmt |
Jarl Dahlerus |
Handelsbanken |
Ulf Gustafsson |
Motorola |
Jörgen Andersson |
Ariad Nordic |
Micael Dahlgren |
Appius |
Birgitta Gustavsson |
Scania |
Leif Andersson |
Försvarets Materielverk |
Göran Danielsson |
Enator Communications |
Leif Göransson |
Mmt |
Rickard Andersson |
Appius |
Anders Danne |
Ericsson Radio Systems |
Elias Hadad |
Janusprojektet |
Thomas Andersson |
Folksam |
Göran Derefeldt |
Sema Group Infodata |
Anna Halling |
Ericsson Consulting |
Gunnar Appehl |
Razorfish |
Carlo Diviani |
Greenwood Communications |
Kristian Hansson |
Asg Road Transport |
Magnus Arnström |
Thoreb Itmobile |
Jan Doverlöv |
Vbb Viak |
Lars Hansson |
Icon Medialab |
Mikael Augustinson |
Aptitude |
Per de Vahl |
Xinit |
Rickard Hansson |
Syntagma |
Samy Awad |
Nokia Mobile Phones |
Björn Ehlert |
Ord & Vetande |
Daniel Hasselberg |
Ellemtel Utvecklings |
Joakim Axelsson |
Enskilda Securities |
Niklas Ekarv |
Ericsson Radio Systems |
Björn Hedenborn |
Asg Road Transport |
Peter Axelsson |
Sj Data |
Kristofer Ekbom |
Ellemtel Utvecklings |
Joakim Hedlund |
Cap Gemini Sverige |
Lars Backhans |
Microsoft |
Camilla Eklund |
Telia Mobile |
Bo Herlin |
Essnet |
Cathrine Bagger |
Ericsson Radio Systems |
Anders Elgemyr |
Fischer Partner Fondkomm |
Fredrik Hesse |
Reterm |
Bjorn Bai |
Ericsson |
Tomas Eliasson |
Sigma Exallon Systems |
Robert Hesselstad |
Mmt |
Ulf Baldhagen |
Tele2 |
Olle Elioth |
Telia Infomedia |
Mats Hjalmarsson |
Ericsson Radio Systems |
Hans Beckérus |
Concis Consulting |
Jonas Elmquist |
Volvo It |
Tina Holmboe |
Icon Medialab |
Fredrik Bekken |
Nutek |
Gunnar Emanuelsson |
Enator Communications |
Lennart Holmgren |
Svenska Dagbladet |
Ulla Bengtson |
Byggindustrin |
Jan J Enejder |
Sigma Exallon Systems |
Charlotte Holmquist |
Ad Vivum |
Anders Berg |
Brg Capital |
Lennart Engelhardt |
Mmt |
Mikael Holmström |
Ericsson Utv |
Mikael Berg |
Intentia R&D |
Johan Engström |
Post- & Telestyrelsen |
Linus Hugo |
Göteborgs-Posten |
Johan Bergelv |
Öhman Fondkomm. |
Kjell Engström |
Sveriges Radio |
Christer Hultgren |
Sema Group |
Robert Berglund |
Wm-Data Teknikkonsult |
Thomas Ericson |
Riksbyggen Ekonomisk |
Anna Hultman |
Ericsson Radio Systems |
Staffan Berglund |
Goldpen Computing |
Mikael Ericsson |
Pacer Communication |
Anders Hultstein |
Hiq Data |
Johan Bergström |
Skogforsk |
Tomas Ericsson |
Wm-Data Teknikkonsult |
Lennart Hägglund |
Ericsson Mobile Comm |
Henrik Bernhold |
Nordström Öhman |
Anders Eriksson |
Riksdagen |
Per Olof Hårsmar |
Smhi |
Hans Bjurström |
Pipebeach |
A-C Eriksson |
Lm Ericsson Data |
Mats Ireneus |
Frontec Tekniksystem |
Lennart Blecher |
Riksdagen |
Marie Eriksson |
Ericsson Utvecklings |
Göran Israelsson |
Enterprise Systems |
Henrik Blomquist |
Slottsbacken |
Peter Eriksson |
Teracom |
Henrik Jangvall |
Callmedia Scandinavia |
Anna-Carin Bodin |
Ericsson |
Magnus Fagerström |
Posten Sverige |
Steen Jensen |
Cap Gemini Sverige |
Mauritz Bomark |
Lantmäteriverket |
Bengt Falke |
Svenska Kommunförbundet |
Erik Johansson |
Telenor Mobil |
Ivan Bretan |
Telia Mobile |
Joachim Feldhoff |
Lm Ericsson Data |
Henrik Johansson |
Tekniska Verken |
Patrik Brinkhäll |
Riksdagen |
Greg Fitzpatrick |
Medianet-Ski |
Jörgen Johansson |
Ericsson Software |
Lars Broman |
Essnet |
Thomas Flink |
Ericsson Radio Messaging |
Magnus Johansson |
Mobileyes |
Johan Bromander |
Ericsson Software Techn |
Anders Forss |
Svenska Spel |
Mikael Johansson |
Ericsson Infotech |
Michael Broms |
Netsolutions |
Marie Friberger |
Gis-Centrum |
Rikard Johnsson |
Ericsson Radio Systems |
Tommy Brundin |
Wm-Data Ellips |
Magnus Fridh |
Skogsstyrelsen |
Håkan Jonsson |
Telia Research |
Ruth Brännvall |
Ericsson Radio Systems |
Robert Fromell |
Wm-Data |
Johan Jonsson |
Scandinavia Online |
Bogdan Bunea |
Linnédata Väst |
Roger Färdigh |
Ab Trav Och Galopp |
Patrick Juthberg |
Teracom |
Per Burdack |
Alcatel |
Lars Gelin |
Essnet |
Sven Jönsson |
Gambro Lundia |
Bengt Bäverman |
Ab Trav Och Galopp |
Karl Gemfeldt |
Region Skåne |
Håkan Karlsson |
Ericsson Networks |
Katarina Bååth |
Hjälpmedelsinstitutet |
Elisabeth Genberg |
Telia Mobile |
Johanna Karlsson |
Telia Promotor |
Mikael Börjeson |
Luleå Tekniska Universitet |
Ingemar Gerbro |
Sema Group |
Jonas Karlsson |
Wide Business Dev |
Dan Cambladh |
Strömgruppen Dc Elkonsult |
Ulf Gerhamn |
Marquardt & Partners |
Roland Kero |
Telia Research |
John Carehag |
Amuze Electronic Media |
Stefan Gidlund |
Wm-Data Infrastruktur |
Jacob Key |
Ariad Nordic |
Curt Klangsell |
Sj Data |
Anders Nygren |
Lm Ericsson Data |
Joakim Sparrentoft |
Optiscan Intermec |
Rolf Kolmodin |
Sun Microsystems |
Jeanette Nylund |
Ericsson Radio Systems |
Lena Spegel |
Tl Information |
Urban Källström |
Sas |
Anette Nyström |
Föreningssparbanken |
Rikard Steiber |
Digiscope |
Michael Lager |
Salcom Communication |
Jenny Nyström |
Cap Gemini |
Joakim Stenberg |
Om Stockholm |
Elisabeth Landén |
Kipling Info Technology |
Elisabet Näslund |
Byggindustrin |
Gösta Steneskog |
Institut V |
Peter Langemar |
Mmt |
Peter Odander |
Ericsson Radio Systems |
Anita Stening |
Sema Group |
Patrik Larsson |
Infocast |
Per Ollas |
Hiq Data |
Mats Stockman |
Birka Service |
Paul Lernmark |
Reston |
Josefine Olofsson |
Communicator Infosys |
Håkan Stolt |
Telia Prosoft |
Jonas Lind |
Telia Infomedia |
Lars Olofsson |
Saab |
Peter Strömbäck |
Teracom |
Anders Lindberg |
L M Ericsson |
Peter Olsson |
Bankgirocentralen Bgc |
Krister Svartsjö |
Ericsson Radio Systems |
Peter Lindblom |
Kentor |
Mary Omming |
Utbildningsradion |
Kennet Svensson |
Lm Ericsson Data |
Thomas Lindbom |
Concis Consulting |
Kjell Ottosson |
Telenordia |
Mona Svensson |
Riksdagen |
Bo Linde |
Microsoft |
Jon Ouzounidis |
Sendit |
Patrik Svensson |
Scania |
Stefan Lindeberg |
Cisco Systems Sweden |
Kjell Tore Overtun |
Ericsson |
Peder Swenman |
Generic Systems |
Pär Lindholm |
Wm-Data Education |
Mikael Palmblad |
Astra Hässle |
Andrew Szelke |
Unwired Planet |
Peter Lindkvist |
Bt Industries |
Mats Parhammar |
Know It Utveckling |
Johan Söderqvist |
Handelsbanken |
Annica Lindmark |
Högskolan I Gävle |
Susanna Paulik |
Telia Mobile |
Henrik Sörling |
Utrikesdepartementet |
Hans Lindqvist |
Enator Communications |
Kricke Perez |
Scandinavia Online |
Voja Tadic |
Arjano Data |
Helena Lindström |
Enterprise Systems |
Carl-Henrik Persson |
Riksdagen |
Henrik Tingwall |
Connecta I2i |
Åke Lindström |
Storsthlms Lokaltrafik |
Roger Persson |
National Semiconductor |
Anna Tjernström |
Europolitan Stores |
Johan Ljungberg |
Ariad Nordic |
Sjöberg Peter |
Ellemtel Utvecklings |
Magnus Torstensson |
Mijada Sverige |
Hans Lund |
Riksdagen |
Åse Petersén |
Ericsson Software |
Malin Trossing |
Cell Consulting Group |
Kent Lundberg |
Nokia Svenska |
Joakim Philipson |
Ariad Nordic |
Jörgen Tränk |
Ericsson Radio Systems |
Per Lundberg |
Sveriges Radio |
Lars Pleiner |
Teracom |
Johan Turén |
SEB |
Stefan Lundgren |
Sendit |
Jan Ragnarsson |
Ericsson Radio Systems |
Mikael Törnborg |
BT Industries |
Lars-Ove Lundh |
Infologigruppen Piteå |
Juha Rapeli |
Philips Consumer Comm. |
Peter Wanger |
Janusprojektet |
Roland Lundqvist |
Telia Telecom |
Roger Reinholdsson |
Sveriges Radio |
Stefan Weil |
SMS Nordic |
Alf Lövbo |
Posten Sverige |
Britt-Marie Rengholt |
SAS |
Jöns Weimarck |
Linnédata Stockholm |
Bo Magnusson |
Ericsson Radio Access |
Pentti Romppanen |
Dialect |
Malin Wemnell |
SEB |
Erik Malmlöf |
Volvo Teknisk Utveckling |
Anders Rosén |
Ericsson Radio Systems |
Richard Widén |
Ericsson Radio Systems |
Mattias Malmnäs |
Netsolutions Sweden |
Karim Ryde |
Wm-Data Education |
Thomas Wignell |
Ericsson Infotech |
Bo Maltanski |
Vägverket |
Lillemor Räntilä |
Sema Group Infodata |
Christer Wilsby |
Sun Microsystems |
Per Merkel |
Gis-Centrum |
Anders Rådell |
Mandator Sverige |
Stefan Wärnsäter |
Posten Sverige |
Johan Mickels |
Hiq Data |
Tony Sajdak |
Referense |
Ulf Änggård |
Ericsson Wireless Internet |
Ethel Milberg |
Sema Group Infodata |
Staffan Sandberg |
Linnédata Väst |
Kent Åberg |
Sun Microsystems |
Björn Molin |
Janusprojektet |
Per-Ove Sassersson |
Handelsbanken |
Thomas Ålund |
Dicom Datautv |
Johan Montelius |
Intelligence |
Björn Schröder |
Saf |
Peter Åström |
Aftonbladet |
Anders Myrén |
Xinit |
Mats Schultze |
Connecta Webtech |
Kristina Ödman |
Wm-Data Teknikkonsult |
Eva Nilsson |
Ericsson Software |
Patrik Servin |
Teracom |
Josef Ögren |
Entra Data |
Fredrik Nilsson |
Communicator Info Syst |
Mohssine Shabraoui |
Ellemtel Utvecklings |
Patrick Österlund |
Lm Ericsson Data |
Johan Nilsson |
Riksdagen |
Maria Sjö |
Naturvårdsverket |
Göran Östlund |
Sveriges Verkstadsindustrier |
Jonas Nilsson |
Sema Group |
Johan Sjöberg |
Microsoft |
||
Krister Nilsson |
Nordström Öhman Wjc |
Otto Sjöberg |
Aftonbladet |
||
Per Nilsson |
Telia Infomedia Partner |
Johan Sjöblom |
Lm Ericsson |
||
Ulla Nilsson |
Electrolux |
Karl-Einar Sjödin |
Nutek |
||
Anders Nordström |
Ericsson Consulting |
Fredrik Sjölander |
Svenska Kommunförbundet |
||
Erik Normann |
Lantmäteriverket |
Veronica Sjöstrand |
Allt Om Stockholm |
||
Patrik Nyberg |
Ericsson Software |
Pär Skogh |
Trav Och Galopp |
||
Edvind Nygaard |
Föreningssparbanken |
Staffan Sköld |
Aspiro |
D. Abma |
Elsevier Science |
Rene Luyk |
ICT WaterProof BV |
Jean-Francois Abramatic |
W3C |
T. Maas |
Stork NV |
P.A. van Akkeren |
Signum |
Jan Massier |
CMG Finance |
Hans Bakker |
A-Net Consulting |
Elizabeth Mattijsen |
xxLINK Internet Services |
Miente Bakker |
CWI |
R.J. van Meenen |
Consultdata Nederland BV |
Rob Bank |
Desk.nl |
Wouter Mettrop |
CWI |
O.A. Beetstra |
PTT Telecom - ICT Innovatie |
K. Metzger |
Philips International |
John Bijlsma |
NNI - NEC |
M. Neerincx |
TNO Technische Menskunde |
Eddy Boeve |
General Design |
J.T.A. Neessen |
TNO Multimedia en Telecommunicatie |
R.L. Bohmer, Jr. |
TNO-TPD |
M.C. van Ooijen |
ServiceTec |
W.J.Th. Boogert |
Rijksuniversiteit Groningen |
Gerard van Oortmerssen |
CWI |
P. Boots |
Audax Tros MultiMedia BV |
Ivana van Overmeeren |
SURFnet ExpertiseCentrum BV |
D. Bouman |
Internet Aware |
Henk de Poot |
Telematica Instituut |
R. Bouwhuis |
Boreas |
P.G.Z. Potgieser |
PTT Telecom MND |
A. Brauers |
Philips Research Laboratories Aachen/Hamburg |
J.Q. la Poutre |
Arthouse Multimedia |
B.A.M. van den Bremer |
Rontgen Technische Dienst BV |
R. Prins |
Industriebank LIOF |
Irma Brevink |
Desk.nl |
Vincent Quint |
W3C |
Judy Brewer |
W3C |
Femke van Raamsdonk |
CWI |
Jan Brinkman |
BRS Premsela Vonk |
M. van Rij |
Elsevier Science |
Theo Brinkman |
School voor de Toekomst |
Marta Rozsa |
CMG Finance |
Remco Brokken |
Internet Aware |
Janne Saarela |
W3C |
Jeroen van Dalen |
Desk.nl |
H. Schaffers |
Telematica Instituut |
Daniel Dardailler |
W3C |
J. Schipper |
Shell International Exploration and Production BV (SIEP) |
Ronald Dekker |
Technische Universiteit Delft |
M.H.R. Schnieders |
Arisco BV |
Josef Dietl |
W3C |
Ben Schouten |
CWI / Desk.nl |
H.J.H. van Dompseler |
Tribute MultiMedia |
Frank Schreuders |
Ediforum |
Patrick Donath |
CMG Finance |
Bart Selders |
SKF Engineering & Research Centre |
Peter van Dorp |
Gorter & Van Dorp Consultancy |
Lean Seuren |
ING Informatie Technologie Centrum |
A.C.M. Dumay |
TNO-FEL |
Remco Sierat |
IBM Nederland NV |
Anne Eijsten |
xxLINK Internet Services |
Anco Smit |
Desk.nl |
Roy van Ettekoven |
Pieter Sanerdamstraat 37bis |
Victor Smit |
CMG Finance |
M.R. Gerritsen |
Van Diepen Van der Kroef Lenior Advocaten |
J. Smits |
STORK Engineers & Contractors |
Ian Ginn |
School voor de Toekomst |
J. Sprey |
Van Diepen Van der Kroef Lenior Advocaten |
Danny Greefhorst |
SERC |
E.J.M. van der Stap |
Baan Company |
R.J. Groen |
IMK Advies |
Jan van der Steen |
CWI |
F.L.J. van Haaften |
Rabobank Nederland |
R. Stierman |
Cistron Internet Services |
Rutger Haagsma |
CMG Finance |
Alex Stil |
Arthouse Multimedia |
D. ter Haar |
Cistron Internet Services |
J.L. Theeuwes |
TNO-TM |
P.A. Haaring |
Rijksinstituut voor Kust en Zee |
Hans van Thiel |
Passeerdersstraat 76 |
Lynda Hardman |
CWI |
Monique van der Veer |
Odina Overheid BV |
Steph Harmon |
Nieuw Amsterdam IT Groep |
Niels Vegter |
Universiteit van Amsterdam |
G.D. Harmsen |
CD-One BV |
H.I.J. Veldhuizen |
Compass Interactive - NedStat |
E. Hassink |
Origin Desktop Business Solutions |
J. Veneman |
Internet Aware |
J. Helberg |
Snow BV |
M.J.J. van der Voort |
HLM Triangle |
Bas Hof |
NetlinQ Groep |
Hans van de Vorst |
Shell International Oil Products |
R.M. Hofstede |
Silicomp Networks BV |
I.F.J. Walraven |
Senter |
Maurice de Hond |
Wegener Arcade |
Eduard Weijgers |
Ediforum |
J.H. Hoogland |
Internet Aware |
Bertram Wesselman |
CMG Den Haag BV |
M.A.W. Houtsma |
Telematica Instituut |
W. van Wezenbeek |
Elsevier Science |
L.L.M.L. Hoyng |
KPN Research |
H. Wieringa |
Elsevier Bedrijfsinformatie BV |
P.H.J. Hupkens |
TNO-FEL |
Karin Wijkmans |
xxLINK Internet Services |
M. van Iterson |
TNO-MET |
Rob Willems |
Origin Desktop Business Solutions |
John Jansen |
Fortis Nederland - IAFN |
Bonnie Williams |
Nieuw Amsterdam IT Groep |
M.J.W. Jeuken |
Shell International Exploration and Production BV (SIEP) |
Bas van Winden |
Arcadis |
C.T.H. de Jong |
DTZ Zadelhoff |
Eckart Wintzen |
Ex'tent |
M.D. Jumelet |
IBM Global Services (Data Sciences) |
Marco Woestenburg |
TNO-FEL |
P.V. Kentie |
Philips Design |
Marjolijn Zegers |
xxLINK Internet Services |
Michiel Kostermans |
Origin Desktop Business Solutions |
M.J. Zielinski |
Euridis - Erasmus Universiteit |
Fons Kuijk |
CWI |
P.H. Zwaard, Jr. |
TNO-FEL |
Frank Kuiper |
CWI |
||
Bas Langelaar |
Coopers & Lybrand |
||
E. Lankhorst |
Kabel-Zaandam BV |
||
Gert Lems |
Ministerie van OCenW / CFI |
||
Maarten Lens |
xxLINK Internet Services |
The objective of this work package was to develop a mirroring system / technology for setting up mirror sites of W3C's web site (www.w3.org) in different locations in Europe. The rationale for installing these mirrors is:
The idea of mirroring the W3C web site was parallel to that of starting a European W3C host, back in 1995. As the Web is the heart of the W3C team, it was essential to have it on all places hosting a W3C development team. The first set of mirrors (one in Grenoble, one in Sophia-Antipolis) had the following features :
The drawbacks of that system were :
The aim of this work package was to enhance this first experiment in order to :
In order to achieve the fixed goals for this work package, we had to solve the following problems :
We tested different methods using different protocols and networks to find the most adequate solution. Details of that experiment were presented at the last review meeting. The conclusions are :
According to these results, we decided to elaborate a solution based on both rsync and satellite. Rsync can handle a small set of mirrors and works fine with well connected countries. We use it to update mirrors housing a satellite uplink. Then, to reach a high number of mirrors and badly connected countries, we use one or multiple satellite uplinks.
You can have a look at the capabilities of the Satellite we used for that work.
Given that mirrors will not be hosted in one of the W3C host sites, we don't want them to store W3C team confidential data. However, a team member should be able to access that data through a mirror. This happens whenever the team member is traveling or in a country that hosts a mirror as all requests use the same host name: www.w3.org. The solution we developed was to transparently proxy such requests to the nearest master sites. This solution required the development of a dedicated module for the Apache HTTP server, the server which we use in both the master and mirror sites.
Error handling is the key for having a robust and reliable service. It does not only involve software mechanism, but also human actions and interventions. We will briefly describe here the software mechanism. You can have look at the Error Handling Scheme page which describes the whole organization.
We will not describe the algorithm we implemented to re-transfer a not received update, and how we ensure the serialization of updates. We will just present the way we handle the two major risks of a mirror :
We distinguish among two classes of failures :
To handle local failures, we developed a "neighbouring" scheme : mirrors are continuously testing their synchronization with master servers. When a mirror server realizes it's out of date, it sends requests to its neighbour mirrors to check if one is synchronized. If it finds one, that means it's a local failure and so the out-of-sync mirror starts transparently proxying all its traffic to its neighbour.
If none of its neighbours are up to date (or if they are unreachable), then the traffic is proxied to the master servers. If the master servers are unreachable too, then the mirror will not proxy the traffic, but will instead serve the info it has and warn the user that the info can be stale since the date of the last update.
This scheme has been implemented directly in the Apache server by means of dedicated modules.
This problem concerns assigning users (W3C customers) to the mirror that's the closest to them, in Internet network distances. We adopted a solution that was developed earlier at an other W3C research project. This solution assign mirror servers to Autonomous System (AS) numbers, rather than to IP network addresses. This assignment is static.
The major problem is to determine if a given Autonomous System (AS) is closer to a particular mirror, in terms of bandwidth. Indeed, we do not have any information about the load of links between AS. In order to find this information, we use a two-step procedure :
To go through the step 1, we are using general databases we get from the three major organizations which are assigning AS numbers and IP network numbers :
For step 2, we developed a special Apache module which logs in ms the time it takes to complete a request.
We fully implemented points 1, 2 and 3. Concerning point 4, we developed the methodology, and validated it. However, some work has to be done to develop some automatic tools to analyze and display HTTP log statistics.
Concerning point 1, in order to test our solution with mixed technologies (rsync + satellite), we started with success 2 mirrors out of the satellite scope, one in Taiwan and the other one in Hong Kong.
Our work in the mirroring activity has been quite innovative and recently we were requested to write an Informational RFC on it. Moreover, one of the Apache modules was put into the public domain and integrated afterwards in an official Apache distribution.
The work done in this activity means that launching a mirror today anywhere in "big" Europe (see the Satellite Scope) will be cheap (less than 5000 EUR) and easy (one man-day). For instance, this concerns the setup of the future offices in Italy, Morocco, and Tunisia. We will soon propose to all W3C members covered by our actual satellite their possibility of hosting a mirror. In other words, we achieved the concept of "mirroring on demand".
The deliverables of this work package were the five mirrors sites running in each W3C office. Today, the situation is :
The question is to know if it was worth it. After three months, we don't have numerical results yet, but we can already say that :
In the short term, we will continue this activity in three directions :
To have a comprehensive view of this work package, you can have a look at
The objective of this workpackage is to show the network performance enhancements, bandwidth usage, caching efficiency, of HTTP/1.1, the latest evolution of the HTTP protocol, used to transport documents on the World Wide Web. The goal is to enhance the semantic of the object sent and to reduce the overhead of connection used to serve those objects.
The work consists in four main parts:
Jigsaw is HTTP/1.1 compliant, according to the latest HTTP/1.1 draft, draft-ietf-http-v11-spec-rev-06.txt, replacing RFC2068. The implementation of HTTP/1.1 in Jigsaw helped in the IETF process by providing implementation and inter-operability testing. The implementation demonstrated the effective gain in terms of bytes and packets sent on the wire.
As a result of Jigsaw's modular design, its HTTP/1.1 protocol stack can be reused in other tools. Amaya-java is taking benefits of this stack, along with the filtering engine which is used to provide caching.
One of the main improvements of HTTP/1.1 is the caching semantic. Jigsaw implements a caching proxy, to demonstrate the full range of HTTP/1.1 improvements, the cacheing can be reused directly by clients using Jigsaw's protocol stack. It provides automatically statistics about its usage and savings (in terms of hits served from cache). Here is an example of auto-generated traces:
| Counter | count | percentage | ||
|---|---|---|---|---|
| Total number of handled requests | 2866 | 100% | ||
| Erred requests | 10 | 0% | ||
| ICP redirects | 0 | 0% | ||
| Non cacheable | 357 | 12% | ||
| Cache Accesses | 2489 | 86% | ||
| Hits (served by cache) | 1108 | 38% | ||
| Hits (revalidations) | 77 | 2% | ||
| Misses (no entry) | 1192 | 41% | ||
| Misses (retrievals) | 112 | 3% | ||
To provide a better way of saving bandwidth, ICP (Internet Cache Protocol) has been implemented. This protocol very light, based on UDP. Its goal is to check in a very fast way that a sibling cache may be able to fulfill the request directly, saving the full download to the remote origin server. Jigsaw implements ICP v2 as a filter on the client module, allowing other client to reuse it in the same way the cache can be. Note that the main goal of ICP was to share information between caches, being able to share information with a client is just a nice side effect.
A multicast extension has been made, based on ICP. The purpose was to replace the local disk caches of browsers by personal caches talking with other personal caches using multicast. This solution provides a better way of sharing resources than the automatic filling of caches using multicast, as it requires n times the disk space of one cache. Also, using multicast solves the main drawback of ICP which is the modification of setup when a new sibling cache is added in the loop. An analysis tool has also been developped to get some statistics about savings.
Every two months, complete releases of Jigsaw, including the reusable client stack, are produced, available from W3C web site. Also, the sources are available directly from W3C public CVS server. Also, developpement versions are available on Jigsaw demo site.
Jigsaw is also distributed as a preconfigured as a caching proxy, known as Jigsaw ProxyPackage. This packaging helps demonstrating the benefits of a real HTTP/1.1 client-side and server-side caching proxy in a real environment.
ICP implementation and its multicast variant has been implemented as client filter, it provides then the same functionalities for all the programs reusing the HTTP stack of Jigsaw. It is bundled with the standard Jigsaw distribution.
Comparisons has been done between HTTP/1.0 and HTTP/1.1, using differents settings for the latest. Those experiments show that savings are importants.
The test setup is based on a synthetized website combining HTML documents and GIF images, based on popular pages. The size of the document set is 42Kb, with images size from 40Kb to 70Kb for a total length of 125Kb. Two types of tests are run, retrieval of documents, then cache revalidation. Three types of network settings are used, describing intranets, high bandwidth internet and modem-like speed.
| Channel | Connection | RTT | MSS |
|---|---|---|---|
| High bandwidth, low latency | LAN - 10Mbit Ethernet | < 1ms |
1460 |
| High bandwidth, high latency | WAN | ~ 90 ms |
1460 |
| Low bandwidth, high latency | PPP - 28.8k modem line | ~ 150 ms |
1460 |
And the results:
| First Time Retrieval | Cache Validation | |||||||
|---|---|---|---|---|---|---|---|---|
Packets |
Bytes |
Sec |
%ov |
Packets |
Bytes |
Sec |
%ov |
|
| HTTP/1.0 | 510.2 |
216289 |
0.97 |
8.6 |
374.8 |
61117 |
0.78 |
19.7 |
| HTTP/1.1 | 281.0 |
191843 |
1.25 |
5.5 |
133.4 |
17694 |
0.89 |
23.2 |
| HTTP/1.1 Pipelined | 181.8 |
191551 |
0.68 |
3.7 |
32.8 |
17694 |
0.54 |
6.9 |
| HTTP/1.1 Pipelined w. compression | 148.8 |
159654 |
0.71 |
3.6 |
32.6 |
17687 |
0.54 |
6.9 |
| First Time Retrieval | Cache Validation | |||||||
|---|---|---|---|---|---|---|---|---|
Packets |
Bytes |
Sec |
%ov |
Packets |
Bytes |
Sec |
%ov |
|
| HTTP/1.0 | 565.8 |
251913 |
4.17 |
8.2 |
389.2 |
62348.0 |
2.96 |
20.0 |
| HTTP/1.1 | 304.0 |
193595 |
6.64 |
5.9 |
137.0 |
18065.6 |
4.95 |
23.3 |
| HTTP/1.1 Pipelined | 214.2 |
193887 |
2.33 |
4.2 |
34.8 |
18233.2 |
1.10 |
7.1 |
| HTTP/1.1 Pipelined w. compression | 183.2 |
161698 |
2.09 |
4.3 |
35.4 |
19102.2 |
1.15 |
6.9 |
| First Time Retrieval | Cache Validation | |||||||
|---|---|---|---|---|---|---|---|---|
Packets |
Bytes |
Sec |
%ov |
Packets |
Bytes |
Sec |
%ov |
|
| HTTP/1.1 | 309.6 |
190687 |
63.8 |
6.1 |
89.2 |
17528 |
12.9 |
16.9 |
| HTTP/1.1 Pipelined | 284.4 |
190735 |
53.3 |
5.6 |
31.0 |
17598 |
5.4 |
6.6 |
| HTTP/1.1 Pipelined w. compression | 234.2 |
159449 |
47.4 |
5.5 |
31.0 |
17591 |
5.4 |
6.6 |
The common use of HTTP/1.1 is HTTP/1.1 with pipeline. This way of operation divide by two the number of packets sent on the wire. When congestion occurs, improvements of only a few percents are enough to lower the load in a significant way, reducing by a factor of two provides a huge saving in network congestion, as packet drops are the main factor of connection slowness.
Another experiment, has been done to test the benefits of HTTP/1.1 over HTTP/1.0 on highly loaded web sites. The tool used has been done by the MISTRAL project of INRIA, and the tests have been done using the latest versions of Jigsaw. The results will be published at WWW8 in Toronto, May 8th 1999. They show that the improvement of HTTP/1.1 is even better when load is high, mainly due to the saving of the connection time.
The multicast shared cache experiment as been done on four people, and the results gathered for more than 3 monthes has shown that an extra 5% were saved with an overhead of 150ms on each request. This low result is explained by the focused area people are working on, which are differents for the four testers. Grouping people by interest have raised the hit rate to 20%.
The goal of the project is to encourage the use of CSS on the Web by: (1) developing a validator that reads a CSS file and checks for correctness; (2) providing a Validator Web service.
All deliverables have been available for some time, and are increasingly being used by the public. E.g., the online validator service has been used more than 9500 times in the month of March 1999.
The common style sheets (renamed to "Core Style Sheets") are a set of eight ready-to-use style sheets providing various color-schemes and layouts. They have been available since the end of 1997, and have undergone several small improvements since then. The designer is Todd Fahrner of Verso. Coordinator for W3C was Håkon Lie. The style sheets are available from the W3C home page or directly from http://www.w3.org/StyleSheets/Core/
The online validation service has also been available since December 1997, at http://jigsaw.w3.org/css-validator/ (and is also listed on the W3C home page.) In May 1998, level 2 of CSS became a W3C Recommendation, and an upgrade made the validator compatible with the new specification.
The online service provides users with three ways to validate a style sheet: they can enter the URL of an existing Web document, they can upload a style sheet to the validator, or they can type a style sheet in a form.
The stand-alone validator can be downloaded and installed by users on their own machine. Once installed, it doesn't need a network connection. The program is ready to run on any platform that supports Java, but the source code is also included. It can be downloaded from http://jigsaw.w3.org/css-validator/DOWNLOAD.html
The following graph shows the "hits" on the validation service until the 23rd of April, 1999. Note that the graph shows the total number of requests to the server, which includes validation requests and downloads of the stand-alone program, but also the HTML pages with documentation and the images that are embedded in the HTML pages. The bar for March represents 9615 validation requests and 696 downloads of the stand-alone validator, plus 50944 requests for other pages and images, for a total of 61255 successful requests. The other 30803 requests were either redirected (the majority) or were invalid.
The online service and the stand-alone validator have been designed and implemented by Philippe Le Hégaret of W3C, and coordinated by Bert Bos, also of W3C.
The objective of this work package first part was to provide an integrated set of PICS (Platform for Internet Content Selection) tools consisting of:
WP5A is part of the Trust Work Package whose role is to demonstrate how to use metadata (such as PICS label) to express statements of trust by some party to be used by another party in an integrated system.
PICS (the Platform for Internet Content Selection), as its name indicates, is a platform of specification, for
These languages and protocols extensions have been W3C Recommendations since 1996/97.
There are commercial implementation available (Microsoft IE, Netscape), and overall Govermental support (EC in particular) for this metadata architecture.
Since PICS has been standardized, the Metadata activity at W3C has been moving toward the use of RDF (Resource Description Framework), a more powerful system similar to PICS in many aspects.
In PICS, labels can be provided by 3 different sources:
The following figure explains how this all fit together.
In this figure, harmful content is available on the INTERNET (at the top) and the USET AGENT (at the botton) wants to filter it.
Between the USER AGENT and the INTERNET sits a PROXY server, which reads a Profile file (in W3C standard PICSRules) and can also access a LABEL BUREAU which stores PICS labels for given pages on the INTERNET (third party rating). The PROXY can also look at embedded label and labels coming from HTTP.
The LABEL BUREAU can implements its storage in different ways, we have been experimenting with both a JDBC (Java Data Base Class) and a flat files format (easier for demo). The labels in this database comes from a LABEL EDITOR which uses a USER AGENT (possibly the same) to establish its rating of individual pages (looking at them and giving a PICS scale value in a particular vocabulary).
The PICSRules database is also created and edited using a specialized EDITOR: a Profile EDITOR.
We note that the USER AGENT can also do its own filtering based on the same LABEL BUREAU and Profile data, without going thru a PROXY.
We have decided early on to make this demonstrator available on the Windows platform primarily (although since it is written in Java, it is portable to other system), and furthermore, on a local (not connected to the Internet system).
This was done to ease the demonstration of the tools in conferences, presentations and other venues where an IP access is not always available and where a Windows machine is always there.
We based all our development on the Java language: we used the Jigsaw HTTP Java server (developped by W3C) as a local server and both the Profile and the Label editor tools were completely written in Java.
For User Agent, we have used Internet Explorer, Amaya and Netscape, for their availibity on the platform.
The first software we wrote for this project was an extension to the Jigsaw server to have it compliant with PICS QUERY Protocol (as defined in PICS Label Distribution specification.
We first implemented the Basic File Based storage format, and then the JDBC/SQL version, which works on Unix with the miniSQL free software server.
The Jigsaw documentation at http://www.w3.org/Jigsaw/Doc/User/pics.html describes completely the configuration for this sample label bureau setting.
This was implemented as separate Java tool which provides a User Interface to an editor of PICS label.
The important features of this tool is that it allows for Browser synchronization and Label Bureau Feed: whenever the user select a new url to rate, a network message is send to the Browser (User Agent), which runs a Java Applet (thus portable) which in turn displays the page/url to be rated.
When the label editing is done, the label can be saved in file or they can be "PUT" to the Label Bureau using an extension to the PICS protocol.
Here's a screen shot of the PICS Label editor:
This is also a separate Java tool, which is used to generate PICSRules compliant profiles.
The important features of this tool is that is allow for both individual PICSRules setting but also support simple white and black lists and a graphical logical language to combine several preference files and white/black list into a complete profile setting.
Here's a screen shot of the PICSRules Profile editor:
It is a server-side filter in Java, attached to the proxyFrame and associated to a LabelBureau.
It reads any W3C Profile language (PICSRules), which means it supports multiple arbitrary PICS rating systems, and uses the GenericAuthFilter of Jigsaw to do access control, and to provide user level configuration.
The following figure shows the different ways Jigsaw can read label to do the filrering:
Labels can come from the Bureau, from the local disk (already cached, or local) and from the HTML META element.
In the absence of true "push" technology, the Intray is kept up to date by repeated refresh requests from the browser.
Clicking on an Activity name opens up a new browser window displaying the Activity page. This page "ActivityPage" takes a parameter indicating the Activity being opened, and requests the necessary data from the Web Server. A servlet retrieves the data from the Workflow Database and packages it as a XML page. This is sent to the browser, where it is unpacked and dynamically inserted into the page, again using DOM facilities.
When the Activity has been completed, the Actor uses the "Done" button to open the results capture popup. This collects the results, packages them as XML and submits them to a servlet on the Web Server. This unpacks them and stores them in the appropriate place in the Workflow Database. It also marks the Activity as completed. The actor has the oportunity to "Suspend" an Activity, in which line "Done" causes interim results to be saved to the Web Server. Unlike "Done", the Activity is not marked as completed.
On completion of an Activity, the Servlet uses the stored process definition to identify the next Activity(s) to be carried out. These will be made available in the appropriate Actor's Intray on the next refresh. Processes definitions are stored in the Database as
This describes the overall structure and operation of the workflow demonstrator.
The demonstrator consists of:
Final packaging of the demonstrator for distribution has not been done, as development is continuing with alternative funding. The latest version of the system can be provided on request, or given the infrastructure requirements for the workflow engine, RAL could provide a "workflow service" for demonstration purposes.
Demonstrations were given at:
Demonstrations/Presentations were also given to several groups of visitors at RAL.
These two work packages demonstrate how collections of translated online documents can be managed, navigated, and linearised with user-given criteria.
The original plan for these two work packages was compiled along the project to have a common goal of:
using external content description i.e. metadata to manage content objects and build value-added services with automated inference processes
The deliverables for these Work Packages are categorized under the following sections:
This management issue can be coped with a specialized RDF schema that introduces the following properties:
We have developed an RDF schema which demonstrates that it is feasible to use RDF to indicate the language and variant information about any Web resource. We have further demonstrated that this information can be easily extracted and understood by different applications. For this purpose, we have developed the RDF parser SiRPAC in Java programming language.
Document in general web collections require more sophisticated management aspects. We have thus further developed the RDF schema to deal with the following aspects:
All these properties of any resource in a Web collection can then be automatically understood by applications
The following figure presents an example setting where Web resources (nodes) have relationships (arcs) with other resource using RDF descriptions
One management aspect often lacking support in Web type of linked document collections is the linearisation of multiple documents into one single document. This can be seen as a reverse operation of splitting one original document into multiple hyperlinked documents.
To manage this linearisation, we have developed a specialized RDF schema that authors of hyperlinked document collections can use. Once an automated process includes ordering information about the position of an individual document within a collection, an application can later unambiguously determine the original ordering.
The following figure presents again how resources can be associated with ordering nodes that imply the original ordering of resources within a Web collection. This seemingly complex figure demonstrates how documents within a Web collection can belong to multiple ordered collections and can be linearised in multiple ways. A more straight-forward implementation would have led to a narrow-minded single linearisation of documents.
The biggest contribution of these work packages has been the SiRPAC - Simple RDF Parser & Compiler. This software was prepared along the RDF specification work which ended only at February 22nd, 1999.
This parser has been made publicly available on the W3C Web site for the general public to download. In addition, the use of this parser and RDF in general has been made easier by introducing a parallel online service where a user can create a new RDF description and see how the parsing process takes place.
These work packages did not deliver an authoring system for RDF metadata but focused more on the available RDF descriptions. This was mainly to demonstrate the usefulness of RDF once it is available and to push the public to create RDF with different approaches.
One reason for no delivering an authoring system were the first attempts to build such as system which appeared too general and thus not directly applicable to any specific domain. We also found that limiting the tool to the vocabularies developed in this work package would not have been useful outside these domains and not for the general public.
The results of these work packages have been placed available on the W3C Web site. The most active development with the SiRPAC parser has led to its placement on the public CVS server. This implies the general public can further improve the parser and log in improvements to the CVS base at the W3C Web site.
This workpackage has explored the use of emerging W3C technologies to collate, maintain, browse and search glossaries of terms. These terms are provided to supply agreed terminology for a particular user domain. By using a standard XML representation of entries in the glossary, an open, hierarchical format can be used to provide linked glossaries (or "hyperglossaries"). These can then be accessed using standard XML tools, and XSL can be utilised to supply presentation details.
The Hyperglossary Worktask has collaborated with the Virtual Hyperglossary Group (VHG) a small UK company promoting the open exchange of information in online glossaries using XML. The VHG have been developing a XML browser called Jumbo, for collating, managing and browsing Hyperglossary information. The Jumbo browser has delivered a demonstrator to the project using three main glossaries from within the biochemical domain: principles of protein structures; extracts from the Institute of Pure and Applied Chemistry Gold Book of chemistry definitions; insulin database from the World Health Organisation, Upsaala Monitoring Centre. These were demonstrated at IST'98 (Wien) late November 1998 and at RSA '98 (London) in December 1998.
Additionally, work has been undertaken to provide end user support for hyperglossaries centred around Internet Explorer 5, which,recently, has enhanced support for XML and XSL. The system developed places a glossary at the server and, using Perl at the server side, allows the user to browse the glossary. Additionally the user can also to query the glossary, the system returning the appropriate entry in response to the user's query based on a keyword. This demonstration uses XSL to format the glossary entry. Thus the demonstration exhibits two of the key emerging recommendations from W3C.
As a demonstration of the utility of the Hyperglossary concept, a partial glossary of W3C terminology has been collated with the assistance of Jenny Raggett. This glossary has been designed to provide explanations of technical W3C terminology to the general, non-specialist user. This has been converted to the Hyperglossary XML format and the user can then interactively query.
With the collaboration of the VHG, a means of representing, and curating linked glossaries has been developed. The utility of this has been demonstrated by providing glossaries in several domains, including bio-chemical information and a glossary of W3C terms. The ability of the user to search this information in conjunction with reading WWW pages has been demonstrated. XSL has been used to provide a user interface onto the glossary for simple querying and presentation.
The capability for multilinguality has been built into the system, although a complete demonstration has not been acheived due to lack of content.
The use of RDF has been planned to provide meta information on the glossaries, that curatorial information on the content. Due to the late arrival of XML enabled browsers (which do not as yet support RDF) coupled with the late delivery of the RDF recommendation this has not been completely realised (alternatively, meta-data is represented and used directly within the XML DTD).
The Hyperglossary worktask has delivered a demonstrator in collaboration with the Virtual Hyperglossary Group (VHG), based around the Jumbo XML browser, and utilising glossaries of biochemical information.
A second demonstrator illustrating how the glossary information could be used and presented to the general user has been developed. This is based around the recently released Internet Explorer 5, and demonstrates the utility of XSL to provide presentation formats for the XML data. This uses a hyperglossary of W3C terms for a general audience.
Demonstrations have been given at:
The goal of this work package was to support the work on the W3C's Synchronized Multimedia Activity. This Activity has focussed on the design of a new language for scheduling multimedia presentations where audio, video, text and graphics are combined in real-time. The language, the Synchronized Multimedia Integration Language (SMIL) is written as an XML application and is currently a W3C Recommendation. Simply put, it enables authors to specify what should be presented when. You can, for example, control the precise time that a sentence is spoken and make it coincide with the display of a given image appearing on the screen.
Experience from both the CD-ROM community and from the Web community suggested that it would be beneficial to adopt a declarative format for expressing media synchronization on the Web as an alternative and complementary approach to scripting languages. Following a workshop in October 1996, W3C established a working group on synchronized multimedia in March 1997. This group focussed on the design of a declarative language and the work gave rise to the SMIL specification in June 1998. Philipp Hoschka chaired the Synchronized Multimedia Working Group - SYMM and was editor of the SMIL specification.
The basic idea is to name media components for text, images, audio and video with URLs and to schedule their presentation either in parallel or in sequence. A typical SMIL presentation has the following characteristics:
The SMIL language has been designed so that it is easy to author simple presentations with a text editor. The key to success for HTML was that attractive hypertext content could be created without requiring a sophisticated authoring tool. The SMIL language achieves the same goal for synchronized hypermedia.
This work was carried out in conjunction with the W3C Working Group where most of the requirements collection took place.
The functionality is specified by the W3C Recommendation on SMIL, which was published in June 1998.
An alpha version of an RTSP server has been developed by W3C, and distributed with a beta version of Jigsaw.
Several alpha versions of SMIL players have been developed by members of the SYMM Working Group, and used for interoperability testing before SMIL became a W3C Recommendation. These tools are:
Numerous SMIL-based products exist today, which are either in beta stage of final products:
SMIL Players
SMIL Authoring Tools
Given that the numerous SMIL-related development efforts outside of W3C-LA indicated that sufficient leverage and industry take-up for this Activity was achieved both within Europe (CWI, RAL, Helio, Oratrix) and outside, the development of a specific W3C-LA codebase was given lower priority after developing the alpha version of the RTSP server. The final release of a synchronisation system was replaced by activities around the integration of Television and the Web. This included organisation of a W3C workshop on "Television and the Web" and the launch of a W3C Interest Group on "Television and the Web".
The W3C-LA demonstrations were conducted with the Grins tool (co-developed by RAL and CWI). Demonstrations of this tool include the following occasions:
The aim of this package was to fully integrate the use of Vector Graphics, and in particular CGM, into Web technology. This included specification development, implementations, demonstrations, and encouragement of vector graphics technology uptake by industry. These aims have all been realised.
Shortly after work was started on a Web Profile for CGM, W3C-LA staff became aware of similar work by the CGM Open Consortium. Many of the members of this group were also members of W3C; liaison meetings were arranged and an agreement reached to work jointly on a single Web Profile for CGM. This ensured that the resulting profile would fully meet industrial needs in addition to providing the desired Web functionality. The graphical facilities were based on proven work in the aerospace industry; linking facilities were derived from W3C work on XML linking.
Modification of presentation by style sheets was found to be neither feasible nor desirable for the target industrial audience. To encourage W3C Member interest in a stylable vector format, RAL and INRIA made a Submission to W3C concerning the inclusion of simple schematics in Web pages. A demonstration implementation was produced by INRIA as part of Amaya and this was shown to considerable interest at the Symposia in Germany and the Netherlands. This resulted in further submissions from W3C Members and, in August 1998, the formation of a Scalable Vector Graphics (SVG) Working Group. Written in XML and stylable using CSS, SVG is aimed at the the artistic, design and advertising communities.
The editor for the W3C Recommendation and the ISO CGM Web Profile were provided by W3C-LA. The Web Profile for CGM was accepted by the W3C Membership and was issued as a W3C Recommendation in January 1999. The chair and two members of the SVG working group was provided by W3C-LA. The SVG specification is currently at Working Draft stage, and will be further developed during 1999 with the aim of reaching W3C Recommendation status
Early on, RAL-CGM was updated to handle the NIST Test Suite and to deglobalise variables so that it could act as a plug-in and the existing Netscape-compatible plug-in was further developed. The RAL CGM Plug-In was then integrated into Amaya as a demonstrator of what could be achieved with a Web Profile.
The Amaya CGM plug-in was demonstrated at EITC in November 1997, CeBit in March 1998, and the W3C-LA Symposia in Stockholm, Bonn and Utrecht. W3C work on Web vector graphics was presented at the London event in December 1998. WebCGM and SVG were described in April 1999 at a tutotial session at Eurographics UK. A public WebCGM interoperability demonstration was held in April 1999 at XML Europe 99 in Granada, involving three industrial implementations of the WebCGM Recommendation.
In general, all the deliverables have been produced except for those where investigation revealed that an alternative approach was more valuable; in such cases the alternative approach was delivered.
A Netscape-compatible plug-in was created which not only works in Amaya but also in compatible browsers such as DocZilla, Opera or Netscape. Following this proof of concept, three WebCGM plug-ins have been developed by industry.
A specification for SVG has been developed and is in its second public draft. CSS-based presentation of SVG was publically demonstrated in April 1999 and further implementations are under development, all provided by W3C Members.
Required functionality was developed, extensively reviewed by W3C Membership and by industry, and formalised in the WebCGM Recommendation. The functionality provides for links from CGM to HTML, HTML to CGM, and within different pictures or view regions in a single CGM file. Distributed symboil libraries, hotspots, and hilighting are also provided for.
The WebCGM profile was developed and became a W3C Recommendation. Precise mechanisms of integration with HTML were developed, using the OBJECT element, which avoided Web page dependency on vendor-specific plug-ins. Interoperable industrial implementations have been demonstrated and others announced. WebCGM is the only vendor-neutral CGM profile with hyperlinking capabilities.
It was originally thought that this would involve conversion from other formats to WebCGM. However, due to the rich hyperlinking and graphical capabilities in both WebCGM and SVG, conversion to these formats is not especially useful. Authoring tools have recently been announced that will directly generate WebCGM. Where previous versions of these tools used vendor-specific hyperlinking mechanisms, the new versions of these tools provide for import of the old CGM files and export in WebCGM.
Because WebCGM is the sole vendor-neutral CGM profile with Web hyperlinking, there is less need to support the interconversion of different profiles; similarly, there is no need to use special server-side programs to determine the profile of CGMs stored on a Web server.
Converters from WebCGM are useful, particularly for printing, and commercial authoring packages widely support the printing of WebCGMs and the creation of raster images of CGMs in formats such as PNG.
This has been very successfull. In April 1999, three freely available WebCGM plug-ins were available from commercial sources (Inso, Intercap, Itedo). New WebCGM versions of existing CGM authoring products were announced, for availability in late summer 1999, from each of these companies. Early adopters included Ematek, who in conjunction with TTA produced an automatic converter for existing wiring diagrams (in the proprietary Autocad DXF format) which generated WebCGM and used the WebCGM hyperlinking facilities for parts lookup and wiring connectivity. Sogitec have a WebCGM-based documentation system that uses WebCGM for intelligent graphics.
Aerospatiale are evaluating WebCGM for documentation delivery, as is Boeing and the US Navy. The Citec DocZilla browser ships with CGM support as standard, to be upgraded to WebCGM support in a future release. Eidon, Infovision, and Infosphere all have content repositories that can import and export CGM and are moving towards WebCGM; Infosphere will also support SVG export.
In additio to these original objectives, other results included work on an ISO standardisation of the PNG raster file format, which has been a W3C Recommendation since October 1996. An agreement was reached with ISO to jointly develop PNG 1.0 as an ISO standard. W3C-LA provided the two editors for this activity. Major editorial revisions were made to the specification to bring it into line with the requirements for ISO/IEC standards. One formal ballot round concluded in December 1998. An editing meeting was held in January 1999 to resolve the comments raised. It transpired that further PNG development has taken place in the PNG Development Group, leading to the release of version 1.1 early in 1999. There was agreement that the ISO/IEC standard should align with version 1.1. Considerable editorial work is being done to achieve this and improve the presentation of the key concepts in PNG. A new document is almost ready for release for a second formal ballot in ISO/IEC and in W3C. WebCGM can contain raster data using the PNG compression method. PNG is now supported in all mainstream Web browsers.
All stated objectives have been met, although in some cases by other means than those originally envisaged.
The demonstration shows the RAL CGM plug-in running in the Amaya browser.
Another demonstration shows the interoperability of three commercial implementations of the WebCGM profile, all displaying graphics and dynamic linking and interaction, using the same set of vendor-neutral Web pages which conform to the WebCGM specifications for the inclusion of CGM into HTML.