Comments on: Should you write your own web framework? http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/ www.thinkhole.org Tue, 02 Feb 2010 16:30:08 +0000 http://wordpress.org/?v=2.9.1 hourly 1 By: John http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/comment-page-1/#comment-49879 John Mon, 02 Apr 2007 15:09:46 +0000 http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/#comment-49879 Heh! You might be right! ;) - John Heh!

You might be right! ;)

- John

]]> By: Joshua Paine http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/comment-page-1/#comment-49389 Joshua Paine Mon, 02 Apr 2007 02:57:18 +0000 http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/#comment-49389 Somebody's been reading Aquinas in between blogs. Somebody’s been reading Aquinas in between blogs.

]]> By: Alexander Botero-Lowry http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/comment-page-1/#comment-49195 Alexander Botero-Lowry Sun, 01 Apr 2007 21:42:39 +0000 http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/#comment-49195 Heh. I just wrote a blog post yesterday about rolling my own framework. I think one of the key points is that a) I understand what's going on at every stage, and b) I get to make the decisions exactly as they best serve me. On the other hand I do spend time working on the framework that could be spent working on the site instead. :) Heh. I just wrote a blog post yesterday about rolling my own framework. I think one of the key points is that a) I understand what’s going on at every stage, and b) I get to make the decisions exactly as they best serve me. On the other hand I do spend time working on the framework that could be spent working on the site instead. :)

]]> By: import this. » Blog Archive » Python and the “Tyranny of Choice” http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/comment-page-1/#comment-187 import this. » Blog Archive » Python and the “Tyranny of Choice” Sat, 11 Mar 2006 01:50:45 +0000 http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/#comment-187 [...] So, sorry Hoang, but I disagree. Having more choices is a Good Thing… even if it makes you think and make decisions. [...] [...] So, sorry Hoang, but I disagree. Having more choices is a Good Thing… even if it makes you think and make decisions. [...]

]]> By: import this. » Blog Archive » Quixote not dead yet http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/comment-page-1/#comment-184 import this. » Blog Archive » Quixote not dead yet Fri, 10 Mar 2006 01:11:59 +0000 http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/#comment-184 [...] Quixote is one of the few frameworks I have haven’t tried, but I personally hope it sticks around. I’m one of those crazy people that thinks competition is a good thing. [...] [...] Quixote is one of the few frameworks I have haven’t tried, but I personally hope it sticks around. I’m one of those crazy people that thinks competition is a good thing. [...]

]]> By: import this. » Blog Archive » Mark Ramm on Python web frameworks http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/comment-page-1/#comment-133 import this. » Blog Archive » Mark Ramm on Python web frameworks Thu, 09 Feb 2006 04:08:02 +0000 http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/#comment-133 [...] Sounds sensible to me. [...] [...] Sounds sensible to me. [...]

]]> By: John http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/comment-page-1/#comment-99 John Tue, 17 Jan 2006 01:45:03 +0000 http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/#comment-99 You definitely make some very good points. I'll be interested to hear which framework(s) you end up choosing in the future (and why). Keep us posted! You definitely make some very good points.

I’ll be interested to hear which framework(s) you end up choosing in the future (and why). Keep us posted!

]]> By: Rick Copeland http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/comment-page-1/#comment-97 Rick Copeland Tue, 17 Jan 2006 00:56:03 +0000 http://thinkhole.org/wp/2006/01/14/should-you-write-your-own-web-framework/#comment-97 I think you have some valid points, but I still have to disagree. But first, I think we can agree that the following two extremes are both bad: * No one ever writes their own framework * Everyone writes their own framework In the first case, everyone's using Zope (not that there's anything wrong with that), and in the second, well, why use a framework at all? The purpose of a framework is, after all, to save development time and give you a stable application with less effort than would be required if you did it from scratch. Now, on to the rebuttals..... #1 - If you're writing a tiny, one-off framework, maybe you can do it faster from scratch than by extending something else out there. But when you're looking at re-implementing 70-80% of something, I think you're kidding yourself if you think the process will be smoother than contributing to an existing project. #2 - Providing people with a *reasonable* number of options is fine; confusing them with dozens of "me, too!" frameworks does end up wasting their time. This is mainly written from personal experience. I read the <a href="http://pyre.third-bit.com/pyweb/" rel="nofollow">PyWebOff</a>. I tried doing some toy projects in Quixote, Zope, and CherryPy. (Twisted intimidates me a bit right now.) I tried to discern why on earth both Subway and TurboGears existed (and I'm glad to see they're merging). Having a few, well-differentiated options is a good thing. I think we're way beyond that by now. (As to the books argument, I think it's a little off, at least if you're talking about fiction. One person's imagination is, I believe, significantly different enough from anothers' to justify sharing with another ;-). On the other hand, if you look at business books, you'll find a few books worth reading and hundreds that do nothing but repeat what the good ones said, poorly.) #3 - OK, you bring up valid points. A lot of what gets released is not focused on the "impact" and "market share" (or "mindshare") of the project. But if you are a True Believer (tm) in the advancement of the technology, I think you'll see it move forward faster if we find more ways of cooperating on projects. There are times when a new framework (or library, etc) is needed. I just don't think it's nearly as often as new frameworks are actually created. Some projects have this fairly well under control. You don't hear too much about libtiff, libc, gcc, zlib, sed, apache, Linux, or samba re-implementations. Maybe it's the complexity, or the pain threshold to create your own. I don't know for certain. But I believe that these projects have benefited greatly from the coordinated effort of their developers. Just my $.02 I think you have some valid points, but I still have to disagree. But first, I think we can agree that the following two extremes are both bad:

* No one ever writes their own framework
* Everyone writes their own framework

In the first case, everyone’s using Zope (not that there’s anything wrong with that), and in the second, well, why use a framework at all? The purpose of a framework is, after all, to save development time and give you a stable application with less effort than would be required if you did it from scratch.

Now, on to the rebuttals…..
#1 – If you’re writing a tiny, one-off framework, maybe you can do it faster from scratch than by extending something else out there. But when you’re looking at re-implementing 70-80% of something, I think you’re kidding yourself if you think the process will be smoother than contributing to an existing project.

#2 – Providing people with a *reasonable* number of options is fine; confusing them with dozens of “me, too!” frameworks does end up wasting their time. This is mainly written from personal experience. I read the PyWebOff. I tried doing some toy projects in Quixote, Zope, and CherryPy. (Twisted intimidates me a bit right now.) I tried to discern why on earth both Subway and TurboGears existed (and I’m glad to see they’re merging). Having a few, well-differentiated options is a good thing. I think we’re way beyond that by now.

(As to the books argument, I think it’s a little off, at least if you’re talking about fiction. One person’s imagination is, I believe, significantly different enough from anothers’ to justify sharing with another ;-). On the other hand, if you look at business books, you’ll find a few books worth reading and hundreds that do nothing but repeat what the good ones said, poorly.)

#3 – OK, you bring up valid points. A lot of what gets released is not focused on the “impact” and “market share” (or “mindshare”) of the project. But if you are a True Believer ™ in the advancement of the technology, I think you’ll see it move forward faster if we find more ways of cooperating on projects.

There are times when a new framework (or library, etc) is needed. I just don’t think it’s nearly as often as new frameworks are actually created. Some projects have this fairly well under control. You don’t hear too much about libtiff, libc, gcc, zlib, sed, apache, Linux, or samba re-implementations. Maybe it’s the complexity, or the pain threshold to create your own. I don’t know for certain. But I believe that these projects have benefited greatly from the coordinated effort of their developers. Just my $.02

]]>