Want to take part in these discussions? Sign in if you have an account, or apply for one below
Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.
I have now disabled caching for the moment; I may turn it back on once the migration is complete, but it was costing unnecessarily much (not really much, but disproportionately much in comparison to everything else) to invalidate things in the cache during the migration. The caching layer is still there and in use as a reverse proxy (handling for example the redirect of the nForum and personal webs) and for caching some things, e.g. historical pages, it is just operating with ’no-cache’ behaviour towards the static site where the nLab pages are located, which means that it does cache things, but always does a check towards the static site as to whether anything has changed. There will probably be a slight slow-down in page viewing due to this, but it should still be snappy.
Re #200 That sounds like a good idea. Almost an nLab hackathon. Perhaps this could also feed into Urs’ desire for his presentation. Give people small pages they can edit and fix small things, or make small improvements, to wet their heels.
Almost an nLab hackathon.
Indeed! Then I’ll probably start a new thread for this purpose once I’ve fixed the table issue, to draw more attention to it and plan it/carry it out.
Re #198 on redirects, there’s also the simple wish to have to type less, even when I know the target name, e.g., if I want to link to a plural.
The hackathon sounds like something overdue, though I imagine there will be quite a few pages needing careful consideration as to whether to delete or modify.
Same problem as 156 and 157 is occurring with torsion group, as torsion group redirects to torsion group.
There appears to be a serious bug in the editing functionality, which results in user edits being discarded: https://nforum.ncatlab.org/discussion/13669/microflexible-sheaf/?Focus=98020#Comment_98020.
John, there’s this thread on the nForum
Hi, Richard! Just wanted to let you know that Urs Schreiber wrote an emotional post on twitter and on what is the nlab page to the effect that ’nlab is dead’. Seems like some kind of miscommunication. Maybe he couldn’t access the forum due to the expired certificate and assumed that you stopped working on migration? Anyway, maybe you could reach out to Urs to sort this out? Thank you for your efforts!
I’m very sorry, but a member of my family fell extremely seriously ill on Sunday, and I have been travelling and without a functioning computer since then. Adeel Khan is extremely kindly looking into the SSL issue.
Unfortunately, whilst several people have been very understanding, I consider the Twitter post and some of the things written on the nLab talk group to be distasteful.
Adeel before me did a lot of great work in getting the nLab stable, and for a number of years I have been building on that, adding functionality and maintaining things behind the scenes. I have been following the nLab from the beginning, and have been a user for most of that time. This year I have been working extremely hard towards a long-term future of the nLab that is highly robust, performant, and stable. All of this with more or less no benefit to myself, just because I felt the project to be worthwhile.
However, it is clear that this work cannot go on without the support of the community. I am not on any social media, and not involved in any online projects except this one, but to me what has been written in the places I mentioned, and the overall attitude/approach, crosses a line which means that I cannot continue. I have always tried to act fairly and with respect to everyone, and have always acted in the way I felt would benefit the nLab most. If I were to have received the backing of the community, or at least not explicit negativity, I would have continued; as it is, I am resigning with immediate effect.
Commenting again to say that I (213) am different from the guest at 211.
I’m currently getting 404 for all nLab pages. Is this part of the certificate issue which Adeel can fix? If not then I beg Richard to try to get things back online before understandably stopping work.
I do understand why Urs is upset. He has been the main contributor to the nLab and was making many edits a day until that was taken away from him without warning. However I do think that some of the things he has said have been alarmist and unwarranted. In particular calling the nLab dead when we were in fact returning to normality. I hope we can find a way to complete the migration and get the nLab into a state where it no longer depends on any one person.
Just a quick note in haste that the 404s started once the nForum came back; I’m not quite sure how that could have happened, but have contacted Adeel. Very grateful to Adeel for bringing the nForum back.
I have reverted ncatlab.org to its December 2021 state, with the old instiki-forked software. The nForum should now be working as well. It may take some time for DNS changes to get propagated everywhere (clearing cache can help).
More details:
The issue with the nForum was just an expired SSL certificate. However, renewing it actually required me to have the DNS pointed back to Saunders instead of the new AWS server, just for our SSL provider LetsEncrypt to be able to verify the renewal request. (When I say “required me to”, that’s just because of my lack of access to / familiarity with AWS.)
After renewing the certificate, I reverted the DNS back to AWS. Somehow, this ended up breaking the nLab itself, and also re-breaking the nForum.
In view of this, and also of Urs’s independent urging, I re-reverted the DNS to Saunders, and re-configured the web server to show the old Instiki-based software.
I don’t have time to do much more right now, but I hope that in the near future, the nLab community can come to a consensus on the future of the nLab software. I think Richard’s new software looks exciting from what I’ve soon and I hope he is willing to reconsider his resignation, as it would be a shame for the tremendous amount of hard work he has done to go to waste. Perhaps we could carry out the migration to the new software in parallel, while the current version continues to exist and be editable? Surely it will not be too much work to, once the new system is ready, take the new edits after Dec 2021 and migrate just those to the new software.
And it looks like the HoTT wiki is now inaccessible, reaching a 404 Not Found page.
Guest #218 and #219:
Is this the wiki you’re referring to? I am able to access it and see edits as recently as today.
Guest #221: if that link is not working for you, please try clearing your cache or using another browser. Is the main nLab also not working for you?
Just a comment from the peanut gallery: I have the utmost respect and admiration for all the hard work, insight, and technical skill that Urs, Richard, and a small army of volunteers have put into this project. I greatly appreciate, and find invaluable, what they have done. I hope that some temporary setbacks in what was no doubt a technically challenging transition, sort of “changing horses in midstream”, do not overly disenhearten and discourage people.
I’d like to echo Keith and add a thank you to Adeel for your work getting the recent downtime fixed!
I don’t have time to do much more right now, but I hope that in the near future, the nLab community can come to a consensus on the future of the nLab software. I think Richard’s new software looks exciting from what I’ve soon and I hope he is willing to reconsider his resignation, as it would be a shame for the tremendous amount of hard work he has done to go to waste. Perhaps we could carry out the migration to the new software in parallel, while the current version continues to exist and be editable? Surely it will not be too much work to, once the new system is ready, take the new edits after Dec 2021 and migrate just those to the new software.
This all sounds perfect if it can be done that way. Do you think the existing server will hold out for the time being?
Thanks Adeel!! What a relief.
I have rented an AWS server since the time we made the first attempt at migration (if I can recover my login data, will have to check, but I am for sure being charged for it each month). If there is a volunteer who is capable and has the time/energy resources to migrate the current installation as is to that server, please contact me.
I don’t know how much money the nLab donations amount to, but I suspect they aren’t actually enough to pay a dev to do full-time work to finish the job Richard started. If this is the case, we should have been advertising the possibility of supporting the nLab financially more heavily, seeking microdonations, perhaps even on a regular basis. Complaining that the money we have is not being spent on a paid developer, and blaming the only active volunteer we have, is disingenuous. Especially when said volunteer has asked more than once for others to join in and help him on a volunteer basis, with apparently little effect.
In any case, as far as our “fund raising drive” went, it was pretty poor from a marketing standpoint. If we are serious about using the goodwill in the category theory to leverage some income, then complaining publicly about the state of the nLab is not how to go about it. We could have pulled a “the nLab needs you!” lever, letting people know that editing is not happening, this is unhealthy long-term, and that hard cash means we could probably pay someone to hurry up the job. I feel Urs’ analogy is flawed, even if I agree it feels like that to him, since for probably 99%+ of users, not much changed about their nLab experience, given how few active editors we have.
Just a very quick message to say thank you very much for all the kind words to those who have taken the time to leave a comment here, much appreciated!
That’s very generous aerskine. It sounds like the top priority is the issue that Urs mentioned in 226, of migrating the nLab as-is to AWS. How difficult do you think it would be to do that?
It’s unfortunate that this drama is playing out publicly, but it’s even more unfortunate that only part of the drama is publicly visible. I don’t think the solution to that is to try to make everything public, but I would like to caution all observers not to leap to conclusions based only on what you can see.
Work is already underway on stabilizing the nLab in both the short and long term. Thanks to everyone who’s offered suggestions and help; we will be sure to reach out to the community if anything is needed.
My wish is to retire quietly from this thread, but I must just record here that I have truly have no idea what insinuations are being made in the previous comment. For my part, I consider everything I have ever done with regard to the nLab to be public knowledge, and there is nothing that I wish to hide or am aware of hiding.
I did not wish to mention it, but I suppose I must now respond specifically to one insinuation that was made in the nLab talk Google Group that I was somehow blocking access to the funds that were donated by the nLab. This is completely untrue: the money has never, at any point, been my responsibility, and I have no idea actually of how to access it. It has always been handled entirely by the Topos Institute; Brendan Fong, who has in the background quietly helped the nLab immensely by putting everything in place for handling these funds and by continuing to oversee them, will be able to confirm this, but there should be no need to make the enquiry.
(The above was me :-). I am by the way happy to share any code with anyone, just ask; the code that I have written over the past years for the nLab, which comprises a large part of the ’old’ software, is completely in the public domain. Just to correct a typo as well: ’by the nLab’ should have read ’for the nLab’.)
Is it possible for you or somebody else to create another repository on the ncatlab Github, and upload the source code over there?
Yes, I can and will do that. I was originally planning to thoroughly document how everything fits together and how to deploy things, etc, and may not have the energy/motivation to do this now, but can at least post the code, and people can get in touch in case there should be any desire to deploy it, etc, in the future. It will probably not be before the end of next week at the earliest unless there is an urgent need for it. The code for the ’old’ nLab, i.e. that which is now live, should be up to date in the git repository.
Why wasn’t this ever put on the main page of the nLab while the migration was happening: reality check
Christian Sattler contacted me to remind me about #239; my apologies for the delay, I have now made everything available here.
Just a note of appreciation for all that Richard has done for the nlab! Thanks in particular for archiving your work. Maybe not farewell but just a pause or break? :-)