Fwd: REST APIs, and Tags

[Max Kanat-Alexander]

On 11/27/2009 02:32 AM, Gervase Markham wrote:
> But one of the problems with keywords is the frozen-in-time-list thing
> that David Ascher proposed. Under this new plan, would keywords get
> unrestricted?

	I'm not sure that they would. Remember, right now you could make
"editbugs" a member of "editkeywords" if you wanted. Why don't you? (The
answer to that question has nearly the same answer as "why not let
everybody make global tags?")

	This is open for debate, but it's something that I'd want more data
from a broad set of organizations on before we really go and make a (for
us) radical decision about it. I'd also be interested to know if other
trackers have implemented a tag system, and what the developers of those
systems think about whether or not their implementation was successful.

> Because if we don't make that split, bugs will start growing enormous
> lists of tags, and people will be disincentivised from adding new ones
> to help them track stuff the way they want to.

	Mmm, I'm not entirely sure that's the case. Everybody can edit the
status whiteboard now, and bugs don't have tremendous numbers of tags in
the Whiteboard. Of course, right now only a small number of people
actually put tags in the status whiteboard, so it may not be the same
once we have a field actually called "tags".

> In particular, if I search for "tag=wibble", then it should show me all
> bugs with either the personal or the global tag "wibble". Search should
> not care about the distinction.

	It won't, but that doesn't require all personal tags to be publicly
visible.

> And yes, it should be possible for someone else to do that search too.
> tag=wibble&taguser=gerv at mozilla.org, or some better UI. So that searches
> can be shared by just sending or posting a URL, without needing any
> complicated internal Bugzilla system.

	That's accomplished most easily, with the present system, by allowing
Bugzilla to have public shared searches.

> Why do the idea of "tags" and the idea of "saved search" have to be
> connected? I've never understood that. How are the currently connected,
> and what was the rationale?

	They don't have to be connected. I think that it was just convenient to
piggyback tags on to Saved Searches, because they essentially do the
same thing, in a way--they list bugs. It would be more database-like to
store a many-to-many mapping of a bug_id and tag_id column for each tag
instead, but shared searches give us footer listing, sharing with a
limited group, etc. for free. I don't think the backend matters as much
as the UI, for this.

	-Max
-- 
http://www.everythingsolved.com/
Competent, Friendly Bugzilla and Perl Services. Everything Else, too.