[Chicago] R: Web Frameworks?
MDiPierro at cti.depaul.edu
Sat Aug 18 01:28:40 CEST 2007
First of all I apologize to all of you if you think I was unfair to your project, in particular SQLAlchemy.
There is almost unanimous consensus that SQLAlchemy has superior functionalities compared to other ORM, including Storm. I am arguing with that. I am discussing what is easier to teach/learn. The point I tried to make is that its API is not as easy to use as the Django one because the mapper introduces additional steps to the user. Wasn't this the reason to make elixir? The documentation is impressive but needs to be organized better.
And have you counted how many clicks it takes from the main SQLAlchemy web page to find one single working example in order to get the idea of what it is?
Anyway, for the last time. It is not my intention to pitch one piece of software over another. I am just saying that here we have one mailing list with excellent people who can write excellent code and yet we have Django, Turbogears, Pylons, CherryPy, SQLAlchemy, SQLObjects, Elixir, Tesla, Kid, Genshi, Mako, etc. etc... almost more names than developers. You know more than I know about them but you cannot expect the average user to know them all and to be able to make a choice. You are the experts. You have to make a choice for them (by forming a committee, vote, etc.). I am talking about branding more than I am talking about technology. There are enough people in this mailing list to produce something of the same quality as J2EE or ASP. You could like to see M$ scared by your product and Sun to port it to Solaris but yet this is not happening.
Why do I care or bother to talk about this? Since recently there was no class at DePaul teaching open source software and no Python. We were 100% Java based. I had to convince my colleagues that it was worth to create such classes and that python is going to be a mainstream language. Nevertheless after a few years I still cannot point anybody to a stable 1.0 python web development framework (except for turbogears but its components are not >=1.0).
Yes I could write my own framework and some of my students could too (Cosmin and Christopher for example) but there no such a thing as "the best framework" unless there is consensus. And it is easier to build consensus if everybody contributes a little bit to the outcome.
I am just trying to motivate you to be more ambitious. I am not here to be the judge of something you have created and you are the expert on.
Imagine a PyCon conference where everybody talks about his own project or imagine the same conference where everybody talks about one aspect of the same large project. That makes news.
Massimo Di Pierro
CTI DePaul University, 243 S Wabash Ave, Chicago, IL 60604
Tel. +1-312-362-5173, Fax. +1-312-362-6116
Email: mdipierro at cs.depaul.edu <mailto:mdipierro at cs.depaul.edu>
Web page: http://www.metacryption.com/mdp/
More information about the Chicago