[web2py:12716] Re: Downside / Critique of web2py

[Timothy Farrell]

1) did you experience any scalability problems with web2py? None, but I foresee some maintenance issues concerning the template system.  This complaint of mine isn't specific to web2py but fundamental to all text-based templating systems.  Try to read some of the templates.  IMHO, they are like reading PERL.  Difficult to understand at a glance.  For this reason, I use Genshi templates to great effect. 2) How large did you scale it? About 10k hits per day so far. 3) What are the features that you believe are missing in web2py and you would like to see in future versions? Before I list mine, I'd like to note that web2py is much more powerful than it comes off as originally.  Like _javascript_, it's so flexible that it's easy to forget what you can do with it.  Some of RobinB's feature requests fall under this.  (As in, you can already do a form of them but it's not obvious or not well documented without buying the manual). I would like to see in web2py: - unittests! web2py already has as certain battery of functional tests, but I would like for it to adopt Storm's mentality of "an untested line of code is a bug." - DAL and HTML helpers that don't rely on an auto-incrementing integer id column - configurable admin availability (like set an ip/netmask to make it available to) - better MVC separation with the HTML helpers (this one is kinda purist of me though) - application inheritance - allow one application to automatically have the models and views of it's "parent" available to it without having to copy-paste models - Quite a few source code cleanups that wouldn't affect the average user. (PEP 8 compliance, spelling fixes, better organization etc.) - More validators - A better, faster editor would be nice.  Editarea, the editor in the Admin interface is slow with large files.  There has been much talk over the Amy editor, but that's slower than Editarea and development on it has all but frozen.  I haven't found my ideal Ajax code editor yet.  Maybe I'll make one. - Use the cgitb module for tracebacks in tickets, they are extremely helpful, but not compatible with the DAL for some reason (I haven't looked into this). Having said that, these are all on my todo list.  As in, given enough time, I will submit patches for all of them.  It's just a matter of time. 4) How does it compare do other web frameworks that you have used?" Frankly, I haven't used any other true frameworks.  I used Cherrypy alone for a while.  But I hit a brick wall when 3.1 came out and my app broke.  The documentation describing the changes in the new version were so meager that I couldn't fix my app easily.  I got on IRC and the guys there were not helpful.  Instead I switched to web2py which has a very nice manual and a very helpful community.  I've also used FormEncode for some of my data validation. That, like Cherrypy, broke compatibility (and my webapp) when it moved to version 1.1.  I have yet to move that code over to using web2py's validators but it will come.  There is a lot to be said for Massimo's commitment to preserving backward compatibility. It seems to me that most of the questions on the list are about using the DAL.  It could be that data abstraction is more complicated than other matters of using web2py, but perhaps the DAL could be reorganized to be more clear for what's going on.  This, of course, would break backward compatibility and therefore must wait for the next major version.  I would like to see some things be more clear like changing db.tablename.drop() to db.drop('tablename') since drop is a database function not a table function.  That's just one example, but now I'm nitpicking. All-in-all, I believe web2py is the best framework out there.  I'll summarize why I didn't go with the other major frameworks: - Django: It's not flexible enough for me.  I'm making a true web-application (not just a blog) and that takes more effort in Django than I think it should. - Turbogears: TG1 is too old with Kid and SQLObject and too hackish if I wanted to use Genshi and SQLAlchemy with it.  TG2 isn't stable enough yet for me to consider it. - Pylons:  Pylons is the Java of Python web frameworks.  There is just too much boilerplate code. - Spyce:  I like/want MVC model development, not the old PHP/ASP-style logic-in-page model. - Roll-your-own from components: too much work, plus I wanted a community to go to rather than 10 communities. Cheers, -tim -- Timothy Farrell <[EMAIL PROTECTED]> Computer Guy Statewide General Insurance Agency (www.swgen.com) --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "web2py Web Framework" group. To post to this group, send email to web2py@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/web2py?hl=en -~----------~----~----~----~------~----~------~--~---