Greg McClure wrote:
>I was curious about the maturity and activity on the Webware project. I have
>a new project that would benefit a great deal from being implemented in
>Python. However a lot of the various frameworks seem to be, well, orphaned
>or not very active. That's a concern because I don't want to depend upon a
>framework that might not be maintained a year from now.
We started using Webware about 2.5 years ago; it seemed the same then,
really. There have been some important improvements since then, but the
core seems so solid that it's not necessary.
We found this group to be exceptionally helpful and expert, and have no
>* Is anyone here using Webware for Web applications involving 50 or more
>concurrent users for a regular, extended span of hours per day?
We don't have any clients that are that big, but we do ALL of our
non-static web work with webware, deploying on unix and windows, using
RDBMS, OODBMS, XMLRPC, all kinds of things. Most of these apps are
deployed on the same server, but I doubt that our concurrency ever gets
above about 10 people.
In general, we've found that webware is never the bottleneck. Sometimes
the database is, or RPC, or CPU-bound tasks like imagemagick, but
webware never seems to get in the way.
>* Are there examples?
Well, as I said, not to that level, but here are some sites that we run
that are all webware:
dalchemy.com foliospace.com naudl.org budl.org ballandsocket.us
peresprojects.com cocktaildb.com errollgarner.com
We've got some extranets that have pretty heavy use, but obviously, I
can't show you those ;-)
>* Has anyone encountered any serious bugs/gotchas with implementing Webware
>PSP pages, or any of the other "kits"? (TaskKit, UserKit, WebKit, WebUtils)
No experience with PSP here (I swore off ASP, and that's just too close
for me). The other kits are sortof moot except for Webkit, and it's very
stable. Once, I had to patch Webkit.Page, to make it handle image-button
submits properly (it worked, but wasn't storing the coordinates). That's
pretty small-fry, and it was easy to fix. AND, it was in the stock for
the next rev.
>Please understand, I'm not trying to cause controversy here, I'm only trying
>to make an informed decision based on others' experience with Webware and
>their perceptions regarding its future. Any information would be helpful.
Summary: I think that you'll find that the software is excellent. The
documentation is poor, but the source comments are good. The conceptual
basis for the software is excellent, too. It helps you get work done
without making you drink any particular water (cough zope cough twisted
cough). It's easy to build layers upon webkit that you find yourself
using over and over again to save you time in your own work... your own
patterns. It's easy to program in python rather than in "webkit"...
which is key, in my opinion. It's also easy to get help.
The biggest problem that we have, I'd say, is lack of releases. There
seems to be pretty interesting work in the CVS branch, but nobody wants
to take responsibility to get releases out. I'm including myself here;
I'm just not up to that particular job, so I'm not complaining. The
problem is more one of perception than of utility; Webkit works great
as-is. But the lack of releases does alienate people.
But lots of us use the non-cvs release for production, paying work. If
you're cool with using CVS, more power to you.
Welcome to the group and the software. I hope that you find what you're