[TIP] PythonTestingToolsTaxonomy: link to Selenium bindings for Python

Mark Sienkiewicz sienkiew at stsci.edu
Thu Aug 25 14:24:09 PDT 2011


Geoff Bache wrote:
> On Wed, Aug 24, 2011 at 4:28 PM, Mark Sienkiewicz <sienkiew at stsci.edu> wrote:
>   
>> I agree, but there is still useful information to be had.  A categorization
>> that comes to mind is:
>>
>> complete + NO active development:  This is what I am looking for.  A tool
>> that works.
>> ...
>>     
>
> I agree this would be great - if we could somehow get and obtain this
> info. Problem is, I'm not going to be downloading 15 tools and
> learning enough about them all to judge "completeness" and nor is
> anyone else I suspect. "Date of last activity" is however relatively
> easy to extract.
>   

The cool thing, though, is that it is on a wiki.  Somebody can fill in 
the values as they happen to notice the state of the various packages.  
A blank entry would also tell you something:  Nobody has looked at this 
one in detail to fill in the value.

So, maybe we have columns "Mature" and "Last modified", and leave them 
both blank until somebody knows what to put there.

Come to think of it, last modified is a maintenance issue itself.  Maybe 
the column name is "inactive" and somebody changes from No to Yes when 
they notice it has not been touched for a year.


>>> Links to source code, even if that is all that is there, is useful for
>>> people who want to develop something and want to check how other
>>> people have already done it.
>>>       
>> The problem is not "source code: yes", but "documentation: no".  So, another
>> categorization:
>>
>> - documentation: good.  I can see what it does.
>>
>> - NO documentation:  I'm not even sure what it does... is it worth spending
>> any time figuring it out?
>>     
>
> I agree. Though it admittedly depends on what we see the purpose of
> the page as. I see the purpose as being a useful resource for people
> looking for usable test tools, rather than an enclyclopaedia of every
> test tool effort there's ever been in Python. Why would anyone want to
> start from undocumented source written in 2003 instead of one of the
> other options?
>   

I agree, but there may be other viewpoints to consider.

Bottom line, though, it is a lot less work to leave an entry on the list 
with "None" in the documentation column than it is to check if there is 
documentation, whether the project actually works, and so on.





More information about the testing-in-python mailing list