Hash: SHA1

Martijn Faassen wrote:
> Martin Aspeli wrote:
>> Peter Bengtsson wrote:
>>>> Philipp has posted a blog entry with a good example of Grok code:
>>>> http://www.z3lab.org/sections/blogs/philipp-weitershausen/2007_01_09_you-thought-zope-3-wasn
>>> Gorgeous!
>>> I especially like that you don't have to manually define the template
>>> since it's picked up automatically by name.
>>> Will the grok effort yield any codegenerating scripts and stuff like
>>> the django folks have?
>>> If not, I have some ideas that I could maybe contribute with at/for
>>> the next sprint even :)
>> Code generation sucks. :)
> I agree. I'll highlight what you say below:
>  > - You should never, ever *need* code generation.
> I think the language/library/framework should be powerful enough to 
> write things down quickly without the need for code generation. Code 
> generation where the end result is something that is expected to be read 
> or modified by humans is generally bad, unless it's just about quickly 
> setting up an empty project. We will be looking into this for Grok.

While I am vehemently opposed to code generation per se, one alternative
which I think is important is to generate stuff *at runtime* from
artifacts which are more understandable to business users than Python
code.  E.g., they might specify schema in a spreadsheet, an HTML form,
or a UML diagram, which we then use to create a schema interface (my
'userschema' package[1] does this already for the first two).

>> But:
>>  - ArchGenXML (hacky though it is) is great for business types because 
>> it takes UML (which business analysts understand and customers can be 
>> talked through) and produces content types they can CRUD with. 
>> Seriously, I've seen people sell big Plone jobs on AGX (kinda scary).
> Yes, and even though I think everything I said above is true, I also 
> think that tools like ArchGenXML can be valuable. I just don't want such 
> a tool to be a way to generate repetitive bad code, as usually I'll be 
> working on the level of the code. If there is to be tools, I want a tool 
> like that that can generate clean non-repetitive code. That is, I don't 
> want a tool to become an excuse to say: oh this code can be repetitive 
> and unwieldy, it's just you use a tool to generate it anyway.
>>  - Don't invent a new code generator. Please. :) PasteScript does quite 
>> well, and seems to be adopted by others, e.g. Pylons. We already use it 
>> to make new egg-like packages for Plone.
> Agreed. We're looking into PasteScript, I believe. Philipp has been 
> looking into this.

[1] http://agendaless.com/Members/tseaver/software/userschema

- --
Tres Seaver          +1 540-429-0999          [EMAIL PROTECTED]
Palladion Software   "Excellence by Design"    http://palladion.com
Version: GnuPG v1.4.2.2 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org


Zope3-dev mailing list
Unsub: http://mail.zope.org/mailman/options/zope3-dev/archive%40mail-archive.com

Reply via email to