Sd2c three Python web frameworks

Other gains from sd2c come from web applications. One is Dong Jing's Gae lecture, and the other is Qian Hongwu's lecture on large-scale Web applications.

Qian Hongwu has a point of view that he must choose the type of technology that has been tested and has rich resources.
Solution. For example, for Web applications, PHP is such a reliable solution. I agree with this. asp is a technology that has failed to pass the test and has been phased out in the past few years.
How is Ms? The reality is cruel. Even Ms gave up ASP itself. A friend of mine has an ASP Website asking me to help find a virtual host space abroad (domestic virtual host providers are too unreliable and the price
As a result, I found four well-known Windows platforms that support ASP.

So I thought of the python I 've been using over the past two years.
Web Framework: Django, turbogears,
Pylons. He was eager to go back to the city that day and failed to wait until the Q & A session of Dong Yan's lecture, so he could not ask his opinion. However, according to Dong Yan's opinion in his lecture
He is very admired. His website "haokan book" was developed with Django-he mentioned this and I remembered seeing him at potato barcamp in.

After returning to Shanghai, I also discussed this topic in the BT Prostitution Group. Ling Hu wrote "What is Django?" To discuss this issue.

Me
The first web development that came into contact with python was the Django introduced by limodou, which has been used since version 0.94. Ling Hu used cherrypy before, so
He prefers turbogears Based on cherrypy. However, after a period of use, I decided to change from Django to TG, because I do not like Django
Some orm and non-XML templates-although this template engine called Jinja is indeed much faster than Tg's default kid. However, I do not like to use tg to configure it by default.
Sqlobject and kid, but like sqlalchemy and genshi. The advantage of TG is that you can choose to use the orm and/or template engine as per your preferences.
As Django makes people feel constrained.

In the past 08 years, when the BT Group wanted to build a new version of the dog skin website, we chose pylons under MK's proposal.
TG can be freely assembled, but the underlying layer is based on Paster rather than cherrypy. I cannot comment on the advantages and disadvantages of the two, but it is certain that pylons has better configuration than Tg.
However, the reason for making such a decision is very simple-Tg2 was no longer based on cherrypy, but to be switched to pylons at the beginning, although Tg2 was not officially launched at that time (now also
Only to beta1 ).

Looking back at the three frameworks after such a hard job, I feel different.

Although Django is an all-inclusive solution, its solution can withstand the test-because it is said that the birth of Django is similar to Ror, it is also created based on actual application development needs. It is a solution similar to Ror, but it is never a counterfeit product.

The idea of turbogears is very good, but because too many components are used and they are very different, the TG combination is reluctant to some extent and has to add a lot of ugly techniques, there are many usage problems.

Although the idea of pylons is similar to TG, it is much more refreshing. The most important thing is that the functions provided by it are not bad at all, such as URL ing (rest required) and support for cache (of course, these Django also have ). It's just a little effort-consuming.

Although I have used TG for some applications, it is a small application with only dozens of users. For large-scale applications, the TG performance is very suspicious. At this point, both Django and pylons are much better.

Return to the "tested" question mentioned above. To talk about the longest-tested solution in the python world, it will probably count on Zope, but that stuff is too heavy, comparable to J2EE.

In addition to Zope, these three solutions (even including ror and so on) cannot be tested over time than PHP, but if you must select one of them, i'm afraid the final winner will be Django.

However, I think pylons will be promising in terms of trends --

Shortly after limodou talked about MerB's unknown-the direction of the uliweb framework, MerB officially merged with ror. If ror is compared to Django, MerB is like pylons. Maybe it is impossible for Django to be added after the formal merger of Tg2 and pylons in the future.