[Zope] Does anyone care whether we deprecate ZClasses?

[Jonathan Cyr]

Big mistake, IMHO,

You have no replacement for intermediate developers... the only decent 
documentation, being the various books, have ZClasses throughout.  
Explain their absence to the newbies grokking for Zope enlightenment.

Not everyone needs/wants development tools, some like the TTW management. 

The ZClass code seems to work flawlessly for me, never a problem.

Who didn't cut their teeth with ZClasses?  ZClasses are the reason I 
realized Zope was so powerful... Gee, how would I use an Object 
database???, the ZClass showed me... and later I added Zope Products to 
the mix with my ZClasses and created a viable business based on Zope.  
How are newbies going to make the leap to the newer Zope offerings 
without a ZClass to start with.

This list is full of experts... and I appreciate the resource... but 
often the intermediate Zope programmer is plowed over... whether it's 
the topic of ZClasses or DTML, or any other way Zope used to do things.  
I have to choose each Zope technology carefully for the time 
investment.  ZClasses were well worth my time measured with the whole 
picture... web design, flash programming, HTML, RSS, javascript, css, 
interactive marketing, systems administration, etc.  I can't say that 
for some Zope technology offerings.

I took the better part of 1 1/2 years developing a business based on 
Zope, heavily relying on ZClasses (and DTML), based on the choices in 
Zope presented at the time.  I bought every book available, which 
pointed to ZClasses and DTML being THE way to go... I bought into the 
TTW strategy of managing web apps, and I like it... ZClasses fit the 
bill for me... I don't need the CMF, APE, ZPT, Plone projects... I need 
a simple way to deliver rich objects in an object database... ZClasses

My business venture, based on Zope is just now starting to bear fruit... 
it would be a shame if the frantic pace of Zope's software lifecycle 
couldn't be patient enough for the real world's business lifecycles. 

I started developing on 2.6.1 and have been able to stay fairly current 
to 2.7.3, and stayed fairly pure, only products added were formulator 
and varimage... figured staying compatible with Zope's core features 
would be to my benefit... ZClasses are a core feature of Zope

Unlike the depreciated Versions feature, which was more of a problem 
that benefit.... ZClasses do work, DTML does, as long as you don't try 
to make them solve complex problems.

Keep ZClasses until there's a replacement for "Simple Rich Objects in 
the Object Database" for the intermediate Zope programmers.

This hits me very close to home, I apologize for any harsh terms and 
appreciate Zope, the Zope team, and the work of all of the folks on this 
list, making it so great.  I hope this helps in the discussion and 
weighing of priorities.

I appreciate you asking, thanks for your time,

Jonathan Cyr
cyrj at cyr.info
WeddingWeblog.com





Jim Fulton wrote:

>
> ZClasses are a feature that support through-the-web development.
> Many people have found them useful in the past, but they have some
> significant deficiencies, including:
>
> - They can't be managed with file-system tools, especially
>   revision control systems like CVS and subversion.
>
> - They don't work well with Python development tools, like
>   profilers and debugger.
>
> - They aren't being actively maintained.
>
> Most serious Zope developers stopped using them a long time
> ago and are frustrated that we still expend resources keeping them
> around.  For example, the release of Zope 2.8 has been delayed
> by the requirement of getting ZClasses working with Zope 2.8.
>
> We could choose to deprecate ZClasses.  If we deprecated them in
> Zope 2.8, they would still work in Zope 2.8 and Zope 2.9, but
> their support would be removed in Zope 2.10.  Would anyone be upset
> if this happened?
>
> Jim
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.zope.org/pipermail/zope/attachments/20050401/51061800/attachment.htm