Some may be thinking "after a decade, isn't it time for another rebuild? Isn't it bad to use old code for so long?"
In reality, Flayrah and its server are regularly updated. Even now, there's support for our version of the site framework, Drupal, thanks to its broad, institution-heavy user base. It can run on more modern versions of PHP – without using any new features - and uses the latest MariaDB, the database that does legwork for search features like the one mentioned here.
There's still room for little tweaks behind the scenes. Long comment chains load faster now that user and profile data is retrieved once per page. Contributor lists and taglines now merge usernames that duplicate usernames in various ways – freeing up space to credit editors.
It'd be possible to add improved group features, with spaces for contributors or editors, as that has an upgrade path; but it's unclear what problem it'd solve (editing collaboration?), and seems overkill compared to simpler solutions. Using time to tag feed items may be more useful.
Some may be thinking "after a decade, isn't it time for another rebuild? Isn't it bad to use old code for so long?"
In reality, Flayrah and its server are regularly updated. Even now, there's support for our version of the site framework, Drupal, thanks to its broad, institution-heavy user base. It can run on more modern versions of PHP – without using any new features - and uses the latest MariaDB, the database that does legwork for search features like the one mentioned here.
At some point that version's security support will end, and we'll have to move on. But it'll be a couple more years yet. Much has changed in web development since 2009; upgrading will be a hassle – and I have other sites to care for (one of which kicked off this round of updates, after I saw Yerf's low-res thumbnails on mobile, and extended the fix to contributor icons) – so I'd want it to last for a while.
As mentioned on Twitter, our best bet for a more modern framework that lasts beyond 2023 is Drupal 10, with CKEditor 5; they plan for June 2022, but they're still pinning down the details. It may be well into 2023 before modules needed to duplicate existing functionality are ready.
There's still room for little tweaks behind the scenes. Long comment chains load faster now that user and profile data is retrieved once per page. Contributor lists and taglines now merge usernames that duplicate usernames in various ways – freeing up space to credit editors.
It'd be possible to add improved group features, with spaces for contributors or editors, as that has an upgrade path; but it's unclear what problem it'd solve (editing collaboration?), and seems overkill compared to simpler solutions. Using time to tag feed items may be more useful.