All posts in the topic Tags (Short link)
Summary
- There are 6 posts — by 4 authors — in this topic.
- Latest post made by Steven Clift at Feb 07 04:23 UTC
I propose that we suppress the feature that enables a user to attach tags to
files uploaded via the web. I suspect that a very small group of users tag
uploaded files. There is currently no way to see the tags that are associated
with a file. The feature therefore has almost no benefit to offset the
usability cost.
I know that the use of tags in GroupServer is controversial, but the following
is clear enough.
1: Although the tags are stored in a database, the only part of the interface
that currently supports tags is on file-upload via the Web.
2: Tagging files or posts via email is problematic.
3: Enabling a user to attach tags to their own posts or files either implies or
is closely related to enabling users to tag other users' posts or files. The
latter can only be done once we have audit trails, so that we can track who
tagged what.
4: Every additional feature complicates the interface, and makes it harder to
understand.
On Thu, 2008-02-07 at 14:24 +1300, Dan Randow wrote:
> 1: Although the tags are stored in a database, the only part of the
> interface that currently supports tags is on file-upload via the Web.
>
> 2: Tagging files or posts via email is problematic.
Very.
I guess my question is this. What makes our tagging 'difficult', while
tagging in other applications (eg. Flickr) it is almost universally
used?
Perhaps we should make tagging files compulsory via the web, rather than
taking it away. Just a thought...
I agree with much of what you're saying, I just don't want to see us
losing something that could be very valuable in the future. Perhaps we
need to make the benefits of tagging more visible (eg. using a tag cloud
somewhere) to encourage people to do more of it?
> …I just don't want to see us losing something that could be very
> valuable in the future.
Dropping the user interface for adding tags is not too painful, as it is
just one page. The core infrastructure could remain ☺
On Thu, 2008-02-07 at 15:25 +1300, Michael JasonSmith wrote:
> > …I just don't want to see us losing something that could be very
> > valuable in the future.
>
> Dropping the user interface for adding tags is not too painful, as it is
> just one page. The core infrastructure could remain ☺
Oh, sorry. I was actually referring to the tags themselves -- if we want
to do something special with tags in the future, we won't have anything
tagged.
I do take the point that they're not being used correctly at the moment
though.
> if we want
> to do something special with tags in the future, we won't have
> anything tagged.
>
> I do take the point that they're not being used correctly at the
> moment though.
With that in mind, I really suspect that we don't have anything tagged
now. We could find out, by analysing the data, of course, but I don't
think it's worth bothering. If we start to render tag clouds, we're just
going further down a path that we know gets difficult fast. (And anyway,
if we make clouds, wouldn't we put data that we do have in them ie
'keywords'?).
We have plenty of priorities that we know will deliver high value.
Leave the existing data, and infrastructure, of course.
When you removed folders I did use tags for 15 or so files that I
uploaded (when you could upload files independently) for a private
group's file library. I also use them in the board@ group. ... I liked
how tags were conceptually folders and more. Of course now you can't
sort files by tags, but you can search for files pretty well.
Why is an e-mail sent with:
tags: fiscal, marketing
in the top line any different than adding tags when posting a file to a
group via the web? How is this problematic compared to scanning for
unsubscribe commands?
I would agree that unless you can do things both via web and e-mail, you
might as well hide it. Adding multiple files added to one post would
allow a different way of clustering files for later access.