As far as I have noticed, there is a fairly new feature of naming pages hierarchically, for example [[books/science/topicXY]].
I do NOT use this feature yet.
BUT, what I have been using for a long time are tags separated by a dot. For example [[BACKLOG.task]].
(and the dot is now the file-level separator for hierarchical tags).
Now I have noticed that additional pages are automatically generated such as [[BACKLOG]] and [[BACKLOG/task]] which I have not created myself.
This is impractical because there could be a potential future conflict with hierarchical tags and I get so many “fake” hits in the search now.
Is this a bug because BACKLOG.task files are mistakenly converted into [[BACKLOG/task]] tags (which is actually the wrong “direction”.)?
Or is it no longer allowed to create tags/pages in LOGseq that contain a dot? (Note: this would be very inconvenient for me - I already have many such tags).
it should be quick and easy to type on the keyboard (in my language)
it should be a valid separator in files (I think this last point is not that important)
The dot was ideal for this.
The space character only works with #-tags if I use [[]] in addition - that would already slow me down when creating the tags.
If necessary, I could possibly switch to /. But I’m not sure if that would be necessary and how I can do it as automatically as possible with so many tags.
For example, I have many tags in the form: #task.team.projectname.related.topic.of.the.task
So adapting that would be a lot of work. I would like to be sure that it is necessary and can stay that way for some time.
… and decided to use the “-” separator and furthermore to both pre&postPend my tags with “-” so as to both make them more visible in a Graph as well as easily searchable in regex queries.
My tags now look like this …
–n-podcasts- >>> meaning [notes | podcasts] (extra “-” in front to “lift” it up alphabetically)
… -cs-km- >>> [computing_software | knowledge_management] -cs-km-logseq- -cs-prog- -cs-prog-python-
… -hn-ntrnts-protein- >>> [health_nutrition | nutrients | protein]
… converting all my tags took a bit of time.
Yesterday I read about the tag-wrangler Obsidian plugin that might be useful; I don’t think we have a similar Logseq plugin yet.
I think I agree with @ichmoimeyo that perhaps - is the best separator to use in your tagging case.
I came from DOS era and never chose to use . as intra name/tags separator because of DOS usage of . as file type separator. That dash - has always been a safe name separator for DOS/Windows. I only started using space as name separator with DOS/Windows long file names and quoting " of names with spaces because they make the keywords more readable. Another safe choice separator in names is underscore _. Other separators tend to have some “unsafe” side-effects.
The advise that gives is applicable for large range of text/tagging software, not just obsidian, I believe …
Spaces are not allowed in tags. So, if you want to have multiple words in a tag, you can use these formats:
Thanks @sabre23t and @ichmoimeyo for your thoughts.
I agree that - would be a good candidate. I also think about changing it using the new namespaces feature with / as delimiter.
But information about “namespace pages” is minimal at the moment (only found this https://logseq.github.io/#/page/namespace%20pages ) and as far as I understand there are still a few features missing to use them as hierarchical tags.
As far as I have found out, it is unproblematic to work with tags that contain dot,
BUT what you should not do → mix them with namespace pages. This can lead to data loss. I have reported the bug here: