[Calendula-devel] re: GNUe analysis - might be helpful

[Darryl Caldwell]

Hi James,

I've been giving GNUe a hard time over on the gnue list these last few
days. Thanks for coming to my house. 

(I've actually been playing with GNUe Designer these last 2 days, since
Savannah has been down. I did follow your suggestion and install
GNUe-Reports and now the wizard in Designer is working as expected.)

I spent 2 weeks with 0.4.2 last year. You have come a long, long way
since then. In my mind it is still questionable whether GNUe is right
for Calendula. I really don't have time to experiment and hope it works
out. I need to scratch my own itch and I need evidence of stability. You
have said that GNUe is run under a needs-based development model and a
IRC-based support model. So are many projects; but I venture that they
all developed good foundations and communities because there was
documentation early on. Look at Remi Delon's CherryPy
(www.cherrypy.org).

I don't want to rely on IRC for your support anymore than I want to rely
on RedHat/SuSE/Mandrake for phone support. Email leaves a better audit
trail, and a better basis for FAQs later on.

If you want wider adoption, I suggest that one or all of the core GNUe
developers stop writing apps for use in production (as you note below)
and make documentation a priority.

Every once in a while a GNUe developer will pop up on a list I am on,
and will say, "Hey, why don't you use GNUe for your project?" I always
complain about the lack of docs and visible signs of use. Show me, don't
tell me. Less outreach, more support, please. Write some short guides,
howtos, updated API docs, etc. You wrote a great email yesterday
describing 4 quick steps to testing Designer. That was wonderful. Now
write something describing how you created the apps you mentioned below.

Other folks on the gnue list have supported my sentiments.

I do really dig the concepts around GNUe. If I don't use it for
Calendula, I can foresee using it for something else down the road.

Thank you, thank you taking the time to join this list and for all of
your support on your gnue list. You can find out more about Calendula at
http://fudosys.com/Calendula.html.

-Darryl

ps> Did I mention that you should write more documentation?

-D



On Wed, 2004-03-17 at 14:22, James Thompson wrote:
> Someone pointed out to me the mail below.  So I figured I'd join and 
> respond.... 
> 
> > Why GNUe isn't good for Calendula:
> > 
> > * Poor documentation
> 
> This is true.  It's worse than I recall it being.  I'll try and address that.
> 
> > * Unstable framework
> 
> I'm not sure which part of GNUe is being referred to here.  I'm using 
> gnue-common based apps 24/7 in local company that does business almost 
> nationwide.  They use gnue-forms based UIs daily to accomplish tasks, such as 
> routing their incoming faxes (billings) to the appropriate customer accounts.
> 
> > * Poor support (very little list traffic), 
> 
> We have gravitated toward IRC based support.  This did make for poor list 
> support.  On the other hand if you catch a developer in IRC you can 
> frequently get patches in minutes working in real time with the developers. 
> 
> > limited adoption 
> 
> This partially falls back on poor documentation.  But any project, such as 
> your own, will also have limited adoption at first.  
> 
> > * inflexible UI design
> 
> I think this is where some of the confusion from the docs is hurting GNUe.  
> All the gnue tools can be used separately.  When using parts of GNUe you are 
> not forced into things like using GNUe-Forms as your UI.  GNUe-POS (available 
> in our svn tree) is based upon gnue-common and wxPython.  I'm currently 
> building a web based project using mod_python, gnue-common, gnue-reports, 
> etc.  I may or may not add gnue-forms based administration forms.  It's 
> basically just another set of tools in your toolbox.
> 
> > 
> > them they said, "well that's how we like to work."
> > 
> 
> This may indeed be what you were told.  If so then I'll apologize and try to 
> explain a little.  We constantly have people people drop by IRC telling us 
> how we're doing things wrong.  We've had people suggest we throw out 10s of 
> thousands of lines of code to do things their way.  Most the time they won't 
> bother to learn our systems before doing so.  Often times these same people 
> are the ones that will never contribute anything.  We've had people come in 
> then quickly fork after telling us we're too slow.  I don't think any single 
> fork has ever survived.  I think they tend to figure out that some of this 
> stuff can't be written fast.  And it's hard to find people willing to write 
> business apps for free in their spare time. 
> 
> Anyway, after a while you tend to get a little touchy when people start to 
> question the hows and whys.  Call it burn out, apathy, whatever you'd like. I 
> know it's not right, but we're all human.  I'm not meaning to say that Darryl 
> said anything wrong.  It's probably more like he caught that grumpy old man 
> GNUe in a bad mood. :)
> 
> So, is GNUe right for your project.  I really can't say as I don't know much 
> about it.  Savannah web seems to be down.  IMHO though, using gnue common 
> alone would save a project an amazing amount of development work.  But I'm a 
> little bias about that :)
> 
> If you have any questions please feel free to ask me.  I'll do what I can to 
> answer them. 
> 
> Take Care,
> James
> _______________________________________________
> Calendula-devel mailing list
> Calendula-devel@fudosys.com
> http://list.fudosys.com/mailman/listinfo/calendula-devel
-- 
Darryl Caldwell darrylc@fudosys.com
Fudo Systems  www.fudosys.com
206/567-5802  "Free Your Computers!"