[Plone-UI] [Plone-developers] Theme Distribution

Alex Clark aclark at aclark.net
Sat Sep 10 04:03:06 UTC 2011


On 9/9/11 10:51 PM, Jon Stahl wrote:
> On Fri, Sep 9, 2011 at 2:48 PM, Ross Patterson<me at rpatterson.net>  wrote:
>> Martin Aspeli<optilude+lists at gmail.com>
>> writes:
>>
>>> On 9 September 2011 00:08, Ross Patterson<me at rpatterson.net>  wrote:
>>>
>>>
>>>      What about versioning themes then?  Wouldn't we have to reinvent a
>>>      bunch
>>>      of wheels if we don't use eggs?
>>>
>>> People know how to version zip files - just append a filename. ;-)
>>>
>>> The point here is that it's completely unreasonable to expect someone
>>> whose skills are in HTML and CSS only to:
>>>
>>>   - create a python package
>>>   - put the right files + manifest in the right places
>>>   - use buildout to create a develop egg
>>>   - upload to pypi
>>>
>>> and then for users of those themes
>>>
>>>   - use buildout to install the egg
>>>
>>> We think it is reasonable for those people to mange zip files through
>>> a control panel on a running Plone site and possibly to extract some
>>> files on the filesystem.
>>
>> Um, I'm asking questions trying to ferret out future complications and
>> interactions, *not* propose we subject themers to what you list above.
>> Let me be clear, I think it would be a bad idea.  I do, however, think
>> the issue of versioning is worth considering a bit more than just
>> dismissing this way.
>>
>> Do themes ever depend on stuff (resources?) from other themes?
>
> I've not seen or heard of such a theme.  I think this would be a
> pretty horrible practice.
>
>> Do themes ever depend on something from another theme that was only added
>> or was removed after a certain point in the release history of that
>> other theme?
>
> Again, I don't think I've ever seen this in the wild, and it is IMHO
> something that should be strongly discouraged.
>
>
>> Do theme consumers/users need to have explicit but
>> somewhat automatic upgrades of themes?  IOW, discovery of new theme
>> releases but also being able to control that explicitly?
>
> I think this is debatable.  My inclination is "probably yes" but then
> again, I can't think of a case where I've seen a theme upgrade that I
> wouldn't want to deploy.
>
>> If any of those answers are yes, then it may be worth it to *use* eggs
>> to distribute themes but wrap the creation, distribution, retrieval, and
>> installation of them such that the user is isolated.  If any of those
>> answers are yes, then this option should certainly be discussed
>> further.

I'm wondering if we should continue this on a single list, e.g UI 
(although maybe a "themes" list would be more appropriate? I'm actually 
just realizing this is a cross post, and I suspect I've been replying to 
one or the other list(s), sorry about that.)

Anyway, I still feel like the real "win" here is to satisfy both use 
cases. I think everyone agrees "zips" == "big end user win". The 
question is how much time/effort do we put into supporting zips inside 
Python packages. My feeling is that people are going to do it anyway 
(e.g. Plone shops), so we should pay some attention to it.


Alex




>
> $0.02,
> jon


-- 
Alex Clark · http://aclark.net



More information about the UI mailing list