Scalar 3?

[alexdryden]

I wanted to put feelers out about the durability of the underlying infrastructure supporting Scalar, and, for those who are managing their own installations, to see if there were any other discussions happening about how to manage continued use.

At my institution, Scalar is a vital part of our service portfolio for the Illinois Open Publishing Network, where I manage the tech stack, and widely used other places around campus. Colleagues on our library infrastructure management team have expressed concern about my use of an application on CI2, and the trouble with the USC instance of Scalar is worrying. I think the difficult reality is that Scalar running on CI2 has an expiration date that may be approaching faster than we'd want.

A lot of my bandwidth over the next 12-18 months will be devoted to another migration project updating other aging web technology in the university library, but we would be very motivated to help port Scalar to another framework (or support another solution) and could commit in advance to a modest weekly schedule, provided there were enough other hands on deck to make success likely.

I know this has come up in these issues at least twice before, and Erik and Craig have said they don't have the bandwidth to manage a port. So, in addition to asking for their input, I'm interested in who else would be able to help.

[alexdryden]

I'm tagging a few folks who have mentioned this issue in the past along with more thoughts of my own. It feels like because CI3 is no longer officially supported, and may not ever receive reliable php 8.2 support from its community maintainer(s), that unless @brucehvn made a great deal of progress towards a CI 3 version, that CI3 may not be the best investment.

Also, a move to CI 4 sounds like a really significant lift. The documentation says that "it is more appropriate to think of converting your app, rather than upgrading it". So, if there is a critical mass of individuals willing to work on moving Scalar into a modern framework, and the amount of work to move to CI 4 is more akin to a migration, then I think it is worth considering if another framework, one with a larger and potentially more reliable institutional partner and support community, might be the best option. This choice may also be influenced by the kinds of experience among those willing to contribute.

@craigdietrich can expand on this, but it sounds like a significant part of the current issue with the USC instance relates to their local infrastructure or other idiosyncratic concerns. So, that is good news in terms of the amount of runway we may have to accomplish this work.

Additionally, I wonder if anyone has other ideas about where we might be able to look for interest in this task.

@ColeDCrawford @paulmer

[ColeDCrawford]

Hi @alexdryden - thanks for the tag. I'm no longer at Harvard on a FT basis, so I don't have much bandwidth to contribute to this. I'm also not much of a PHP developer, mostly Python and Javascript. But I agree with everything you're saying. Harvard is still running a Scalar install as a service for affiliates, so we haven't been directly impacted by the USC service instability. @arthurian is probably the best contact for that team.

[ColeDCrawford]

Choosing a different framework entirely would be an interesting development. Laravel seems to be the current leader for PHP development if you're trying to reuse as much logic as possible. Python and Django are probably the language and framework that are most widely used in the DH space if you are considering a wholesale switch.

[alexdryden]

Thanks @ColeDCrawford -- that is a really interesting point about the use of Python/Django among DH folks. There is certainly a model for this kind of work where the conceptual mapping into the new framework and code review is done by a few experienced developers, and migrating smaller chunks of logic can be done by community users who have some programming experience--most likely DH folks with some Python experience. That might also have the virtue of not requiring as significant of an upfront commitment from full-time developers and allowing for a more flexibility.

@arthurian I welcome any thoughts you have, or if you can think of someone else who might want to contribute to the discussion that would also be great.

[ColeDCrawford]

Given NEH and IMLS evisceration, the timing for a major upgrade or rewrite is pretty rough unless you can secure a big chunk of institutional or private funding.

[alexdryden]

Yeah, it definitely makes things more difficult. And even some of the private funding sources are going to be more spartan as those groups divert funds typically earmarked for this kind of work into projects that have suddenly lost their budgets. I still think that it is worth seeing how much capacity is out in the community. Not that we could "stone soup" the whole thing, but that any amount of momentum becomes something to build on. And as funding streams tighten, we might see a growing appetite for funding projects that have a certain degree of bootstrapping.

[brucehvn]

For some reason my response to Alex didn't make it into the issue here. I'm pasting it in now.

Hi Alex,

I'm the person who has been working with Craig and Erik for a couple years now to help maintain the scalar server. We have been faced with the same issues you have brought up. Not only the fact that Scalar is running on CI2, but our server software, OS, etc. is severely out of date.

I have been working on upgrading Scalar to CI3 for quite some time. We first converted one of our dev instances to it and it ran great for over a year. As of just this past weekend, I converted the works instance to CI3.

But since we upgraded to PHP 8, we've had other issues that didn't get resolved just by upgrading CodeIgniter. We're going to try and spin up a new server with an up-to-date OS and software and hopefully that will run better. We appreciate your patience. It's been a nightmare all around not only for customers, but for us as well.

Sincerely,
Bruce Hellstrom

[alexdryden]

@brucehvn that is really great to hear! Can you point us to the repo with the CI3 version? I'd love to spin it up on our hardware.

[arthurian]

We manage our own installation of Scalar at Harvard as @ColeDCrawford mentioned and are also concerned about the situation with CI2. A few years ago, we considered attempting to upgrade to CI3, but unfortunately we weren't able to allocate sufficient resources to do that, so we never got it off the ground. We'd be interested to see the CI3 version @brucehvn and could kick the tires in our dev environment. In general, we're more of a python shop (mostly django) and our colleagues in DH also work in python, so if there was an initiative to migrate to a modern framework, that might be a good target to broaden the base of developers who could contribute.

[eloyer]

Appreciate the offers to kick the tires, it'll be very helpful — we're not quite ready to have folks testing just yet, but will keep you apprised.