User blog comment:MisterWoodhouse/Pausing UCP migrations to resolve some critical blockers/@comment-3218221-20200902000444


 * "Even with that, nothing can compare with a product being out in the wild for us to know how it's being received. The Unified Community Platform has been live on new wikis for a while, but seeing well-established communities on the UCP and hearing your feedback, which was no longer abstract, was vital."

Pausing migrations is the right decision, I think. I'd been experimenting with UCP on your (Woodhouse's) UCP sandbox in advance of batch migrations in a bid to get used to it, but there's obviously such a stark difference between experimenting with an editor on a test wiki and handling the ubiquitous effects of a migration across the whole of a well-established wiki.

The oldest and most popular wiki I admin (504 pages as of September 1, 2020) was migrated a little over 10 days ago. I had less time to prepare for the migration than I thought; I thought we were supposed to have at least a fortnight to prepare for the migration starting from the announcement banner's first appearance, but the wiki was migrated in less than two weeks. The homepage polls were broken (since UCP doesn't support polls), article infobox tabs made wonky, .ogg player does not display, our article background broke on account of FANDOM quietly changing class selectors via the upgrade...I'd not realized the class selectors would change. Did not expect TOC to suddenly reflect typeface changes to section headers. So many little changes...

I'd been following along with UCP news, I've kept my eye on the user-generated list of UCP issues, I experiment on your sandbox wiki...and it goes to show nothing could've quite prepared me for the migration of a 500+ page wiki. It's not entirely easy to adjust to both for users and for the wiki itself. I assume that old Blog (or was it a Help) page about ImageServing (how thumbnail images are selected) must be deprecated, now, since the thumbnail images have been going haywire in category pages. (On that note, I notice that the migration also quietly removed trending articles from the Classic Category view? It's still vastly superior to the Modern Category view, but that removal puzzled me.)

I certainly hope the 2010 editor will provide a faster experience than that which the UCP editor has provided so far. I found the process of editing a long article in Source Editor to be frustratingly tedious, having to frequently wait for what I was typing to appear on-screen, having the editor pause when I wanted to scroll... There was an issue where I foolishly switched from Source Editor to Visual Editor and tried to switch back to Source Editor, but the page froze. Even short articles can experience a delay between input and output.

(It would also be nice if the 2010 editor retains the "predict/suggest article title" ability when a user is preparing to hyperlink to another article. UCP Source Editor seems to have dropped this helpful function, causing a user to potentially have to look up the long article title they don't remember off the top of their heads. Ah, could 'Link Suggest' be referring to that ability?)


 * "it leads users to believe content was lost in migration,"

To be fair, some content has, hasn't it? Like the two polls on 'my' migrated wiki's homepage that went defunct. Perhaps the migration banner could provide more than one link? (e.g. one to a useful blog post, another to the UCP help page.)