[Twisted-web] status of Twisted Web and Web 2

[Andrew McNabb]

On Thu, Mar 06, 2008 at 02:18:28AM -0000, glyph at divmod.com wrote:
>
> You should read
> http://twistedmatrix.com/trac/wiki/WebDevelopmentWithTwisted if you
> haven't already.

Yes, I've read that, but it's pretty general rather than detailed.  For
example, when I read "I want to write a REST / WS-* interface to my
Twisted application: In this case your best bet is to use Twisted Web
directly. Porting to Twisted Web 2 for these kinds of applications will
be fairly trivial (changing a few imports, method names)."  I remember
reading this a long time ago, and it wasn't clear whether anything had
changed recently or how distant these changes were.

>> Is Twisted Web 2 close to becoming stable?
>
> Not unless a lot of qualified people discover a lot of free time in the 
> near future.
>> Is it an active project (most of the tickets I saw on Trac were a year or 
>> two old)?
>
> Yes, sort of.  There really isn't a lot of manpower available.  But it's 
> not dead.

That's not the most reassuring answer I could have expected. :)

>> How is its support for HTTP 1.1?
>
> It is an HTTP 1.1 server.  What else do you mean?

Stability and interoperability were the main things I had in mind.  How
well have the server and client been tested against other clients and
servers out there?  Does it make it easy to deal with persistent
connections?

>> How complete is XMLRPC support on both client and server?
>
> I don't personally know, but isn't XMLRPC kind of trivially simple by 
> design?  I've never heard of anyone having difficulty with XMLRPC support 
> in Twisted.

It is indeed simple, but just because something is simple doesn't mean
it works well. :)  Your second statement answers the question, though.

>> Is there anything that Twisted Web is still better at than Web 2?
>
> Twisted Web is still quite a bit faster.  It's also simpler.  It might have 
> fewer bugs, just by virtue of having been around longer and being more 
> widely used.  On the other hand it might have more bugs because there has 
> been less enthusiasm to work on it.

That is very useful information.

>> I've written some code that uses Twisted Web, and it seemed a little
>> basic.  For example, I had to make some modifications in order to
>> receive large files in progress instead of waiting until they were done
>> downloading.
>
> It would be nice if you would contribute those modifications as a patch and 
> sheperd it through the process of being merged.  If more people did that 
> perhaps the project wouldn't be so stagnant :).

It's a chicken and egg sort of thing.  When things seem quiet, you're
not sure if your patch will be appreciated. :)

In general, I'd be happy to submit this as a patch.  There's a change
I'd have to make first, though.  I wouldn't want to be ashamed of my
submission.  I'll bring that up in a new thread.

>> I have a new project that would use a lot of HTTP and XMLRPC, and I was
>> wondering if it would be safer to go with Twisted Web or Twisted Web 2.
>
> I doubt you'll have any serious problems with either one.  Unfortunately 
> there is no clearly correct answer to your question at this point, but if 
> you make the "wrong" decision, it shouldn't be terribly difficult to port 
> your code to the "right" one later.

That's very helpful information.

>> I was also wondering if there are any recent or forthcoming advancements
>> to look forward to.
>
> I would avoid anticipating any work on the part of the Twisted team.  I 
> think that it's safe to say we are universally overworked.  There's lots of 
> stuff we'd *like* to do, but not a lot that we're really able to :).

I know the feeling. :)

> I'd also like to fold in large parts of Nevow and get rid, at the very 
> least, of Nevow's application server components.

From my perspective, the most compelling thing about Twisted is the
ability to use it within other programs.  I find the low-level protocols
more uniquely useful than higher-level frameworks, especially in the
face of competition with large projects like Django.

Thank you very much for your feedback.  My questions were intentionally
open-ended, and you did a fine job answering them.  Thanks.

-- 
Andrew McNabb
http://www.mcnabbs.org/andrew/
PGP Fingerprint: 8A17 B57C 6879 1863 DE55  8012 AB4D 6098 8826 6868
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://twistedmatrix.com/pipermail/twisted-web/attachments/20080305/af7893dd/attachment.pgp