Thanks! Yes, that definitely helps to explain your motivation. It's a good example of why decentralisation and open standards are so much harder than having a single monolithic app stack.
We had a hope that Solid 0.9 would be a "black and white" contract between app developers and server developers, and we had to conclude that this hope was too early. But over the years to come, Solid will get better and better, just like the web platform did after the days of IE6.
And in the meantime we will keep maintaining https://github.com/solid/test-suite#table like Solid's own equivalent of https://caniuse.com/ so that app developers know what to expect from each different server!
Thanks again for the above posts @RubenVerborgh , @michielbdejong and @sjoertrix:utwente.io! They convey the constructive atmosphere of our meeting yesterday on specs vs implementations. Please see Participants and all the links in https://indico.cern.ch/event/1124784/. Pasting here the summary. Please place here comments/corrections/additions.
The Solid specification being now on version 0.9.0, meaning close to final, further changes can only be few, well justified and agreed.
The top-of-the-list Solid server implementations NSS, ESS, CSS show deviations from the specs on few specific points (see table).
These deviations were well thought of and done by justifiable design choices and to serve existing use cases.
Adherence to the specs is not a dogmatic issue but a technical one.
The test-suite, users' pods migration, ACLs re-adjustment, Solid applications' development are affected by the differences amongst Solid server implementations and get discouraged in the process of getting familiar with the Solid technical solutions and embrace the Solid project ideology. This is why the necessary changes to the specs should continue to be discussed between editors and implementors now that we are at an important point for Solid.
This was agreed by all participants and started immediately by with this example issue discussed at the meeting . Mutual thought and talk shall continue with ad-hoc technical meetings between all interested parties, i.e. the Solid project leadership, specs' editors, implementors, apps' developers, "customers". Posting comments in github issues is sometimes not enough to fully understand the various positions.
Next technical event is https://indico.cern.ch/event/1124804/ on CSS.
@RubenVerborgh @joachimvh @bourgeoa @joeitu @janschill :
You were present yesterday at the above mentioned meeting. Here are the notes. Please add/comment/correct!
The very executive summary is:
These 3 points are really much too brief. Please click and read.
@all Solid World this Thursday! Please see on https://solidproject.org/events how to join!
The Web is 33 years' old this Saturday March 12th! Right?
I also take the opportunity to remind you of Solid World this Thursday 14 April at 16h00 CEST. Agenda and Zoom coordinates on https://www.eventbrite.co.uk/e/solid-world-2022-tickets-223005725127
Notes from last week's Solid World in https://indico.cern.ch/event/1151484/ - credit is due to @KyraAssaad for collecting and sorting the links from the Zoom chat. Thanks Kyra!
Dear @/all ,
the above set of slides describe the new CERN IT strategy, operational model and structure that becomes effective today.
The CERN-Solid collaboration project is discontinued. Despite my efforts, no way to reverse this decision for now. This channel and http://solid.cern.ch will continue existing.
I'll be available for @joeitu until he graduates and shall keep the UI/UX meeting we have arranged with Sharon, Timea and @janschill .
More CERN Academic Training on Linked Data and Solid is being prepared for the autumn with @RubenVerborgh and Uni Gent collaborators.
Many thanks to @timbl , @michielbdejong , @janschill , @RubenVerborgh , @csarven , @pferreir , @ThiefMaster and all the people from Sweden, Austria, Belgium, the Netherlands and Switzerland that helped, supported, exchanged, brainstormed in this effort.
Please find below details related to the forthcoming Academic Training Lecture series that will take place this week on 11-13 May at 16h00 CEST.
You may follow the lectures via Webcast.
There will also be publicly accessible recording.
We shall have the honour to have Internet father Vint Cerf and Web inventor Sir @timbl joining part of this series.
Title: Distributed computing - A historical perspective
After the ARPANET (Advanced Research Projects Agency Network) began in 1969, leading to the TCP/IP protocol development in the USA, a period of intense network competition began between continents and companies. It lasted for more than a decade and affected many protocols and applications. A number of EU-funded projects were born and have kept re-incarnating until today. How was email done in the early 1990’s? When could we send photos by email for the first time? What was the first decade of the World Wide Web like? How did HEP move from mainframes to distributed computing? Which manufacturers have since dominated the landscape in the process towards today’s computing cloud? In this series, we shall try to answer such questions, with contributions by some of the makers and leaders of this technology.
Wednesday May 11th at 16:00 CEST on Zoom https://indico.cern.ch/event/1052073/ - The Network(s)
Ben Segal will explain how trying to solve a straightforward problem - sharing computer data at CERN in the 1980’s - involved us unexpectedly with economic, political and social issues. A study of disruption at CERN and beyond.
Thursday May 12th at 16:00 CEST on Zoom https://indico.cern.ch/event/1052076/ - Email and the Web
Maria Dimou & Nathaniel Borenstein will describe how the protocol wars affected the email, probably the most popular application of the 1980ies. How the Multi-Purpose Internet Mail Extensions (MIME) introduced more than plaintext in email messages. And, last but not least, the Web at CERN during the first decade of its life.
Friday May 13th at 16:00 CEST on Zoom https://indico.cern.ch/event/1052077/ - The Shift and Grid projects
Frédéric Hemmer & Les Robertson will speak about the move from mainframes to clusters, from many manufacturers (HP, Digital, Silicon Graphics and more) to Linux "pizza" boxes, after the "experiment" to run physics code on Windows, the purpose of RFIO, the Shift and Grid projects.
As sponsor and co-speaker in this series, I am very sorry that my lecture clashes with the next Solid World but the CERN Academic Training 2022-2023 programme was fixed before Solid World moved to the middle of the month.
FYI, new CERN hiring programme for young professionals, many IT jobs, if you are interested or wish to relay to your contacts:
Early Career Professionals – University Graduates and the potential projects can be viewed here https://careers.cern/origin-university-graduates
Early Career Professionals – Technician/administrative fields and the projects can be viewed here https://careers.cern/origin-technicians-admins