In my case, a TW with only 30k entries slowed to a crawl when using the tag filter. This was AFTER indexing had been added. Most of the tiddlers in the TW had the same tag. My workaround was to change my filters to use search filter instead of tag filter. As best as I understand it, the indexing is set up to look for a small subset of tiddlers with a certain tag, not a large number of tiddlers with the same tag.
In terms of user management, the problem is that you can have tags that are meta data, tags that are semantic, and tags that are functional. My solution is to have a selectable filter in the editor which changes which set of tags that I see. This could be expanded in other ways. For instance, if you wanted to only see tags that were also contacts, for labeling correspondence. On Wednesday, July 14, 2021 at 9:22:33 AM UTC-7 PMario wrote: > On Wednesday, July 14, 2021 at 5:29:38 PM UTC+2 Stobot wrote: > > I agree with most of the points of where fields are more useful than tags >> as you get deeper and deeper into organization, and the methods make sense, >> but often there's talk about a performance implication that I don't >> understand, but am hoping to as I'm kind of a performance junkie. >> > > Field values and > Tag values are internally indexed since TW version 5.1.20 > > Backlinks are indexed since 5.1.22 > > See: https://tiddlywiki.com/#Release%205.1.20 and search for "indexer" to > see the release message and the link to the PR in the repo. > > >> So my questions are: >> 1. Is the "bad" type of "tag pollution" related to number of unique tags, >> or number of tiddlers *with* tags? >> > > At the moment it should only be a "user management" problem, not a > performance problem any more. > > >> Or both? For example I use my main TiddlyWiki for project organization >> and most details are in fields, but almost every tiddler is tagged as >> either a task, project, meeting, etc. So, I have a small number of unique >> tags, but high number of tagged items. >> > > Filters should be reasonably performant for your use-case > > 2. What is the source of the associated negative performance impact? >> > > It was the number of _all_ tiddlers. As Jeremy mentioned there in the > release-note with 60 000 tiddlers the refresh time improved by the factor > of 3. > Refresh time will matter, if you eg: Switch tabs ... > > >> I assume it has something to do with indexing - and if so there's >> probably a speed vs. memory tradeoff. >> > > Yes, ... We trade more memory consumption for better speed _now_ > > -mario > > -- You received this message because you are subscribed to the Google Groups "TiddlyWiki" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/tiddlywiki/d3cf2375-ce19-443b-924d-552153c91991n%40googlegroups.com.
