Tentative roadmap for bzr web interfaces

[Martin Albisetti]

On Thu, Apr 24, 2008 at 3:57 AM, Jonathan Lange <jml at mumak.net> wrote:
>  I'm curious as to why you want to do this, and why you want to do it first.
>
>  What are the benefits of such an interface over simply improving Loggerhead?

The current interface is fairly flat, and, to do things properly in
the new "web 2.0" world, some significant changes have to be made on
both backend and frontend.
So, considering the web interface, if done properly, can be used by
any backend, it seems like a no-brainer to me to do things right from
the beginning, and build on top for that.

Just improving Loggerhead would be good, but with some extra work, it
can pave the road for other usages for the web interface.



>  Robert had some rather good ideas on this in an earlier mailing list
>  thread "loggerhead navigation".[1]

Yes, that plus what was discussed in the sprint is what's going on in me head.


>  Looms are another thing that would benefit from a top-notch web interface.

I'm still scared of looms, but I suppose once the interface is done
and pluggable, I can try and figure the remaining bits out.


>  Plugins for the win.
>
>  Exporting plain HTML isn't something I think I've heard anyone want.
>  Certainly it would be good for the UI to work without Javascript.

This one is tricky.
I'll look into it a bit more.


>  This reminds me: the web UI should make it really easy to point random
>  people at files. I've often used Launchpad's loggerhead as an easy way
>  to show people a nifty thing that bzrlib or Twisted does that they
>  should copy. It's unfortunate that loggerhead makes these URLs
>  cumbersome.

I promise it will  :)


>  >  == Make installing easy ==
>  >  Currently, installing loggerhead is a real pain, and there are quite a
>  >  few hoops to jump through.
>  >  I haven't looked into it, but it might be worth making it not depend
>  >  explicitly on python 2.4 (I'm not sure why it does, maybe Michael can
>  >  shed some light on this), and build a deb for it.
>  >  It might get more people using it, consequently more chances of
>  >  getting contributors.
>  >
>
>  I personally think that this should be the first thing anyone fixes.
>  Hard install means no users means no developers means no improvements.

I might agree with this, although I really want a better Loggerhead
with a new interface before an easy to install one. But that may be a
more personal choice then what is generally better.


>  Well, we already have a light-weight HTTP server. It's used in tests
>  for the HTTP transport, among other things.
>
>  Do you mean a light-weight webapp? If so, the interesting thing isn't
>  so much the HTTP server as the templating toolkit.

Yes, I didn't expres myself correctly.
light-weight webapp  :)



>  >  == Integrating the new Interface ==
>  >  Loggerhead will need quite a lot of work to use the new interface, so
>  >  a big chunk of time will have to be spent re-working bits of the
>  >  backend.
>
>  Or you could derive your interface from the code in Loggerhead,
>  incrementally improving and extracting Loggerhead's code until you
>  have what you want.

Not if I'm going to do it the way I intend to, but I can try and find
a middle ground.


>  Let me say again that I am very glad that you are looking into this. I
>  can't wait to see the results.

Thanks for all the input :)



Martin