OK. Philosophy time again. If I write it with Python as the day to day interface -- even for some parts of it -- then I alienate the Perl guys and the Java guys and the....
As some of you know, I *LOVE* python. But I gave up trying to force others to love it. There are some hardcore Perl geeks out there that like Puffin but hate Python. I have even more requests for a Perl-based scripting interface than I do a Python-based one. Now what you suggest about making the underlying data structure Python does make a lot of sense. I will continue to look and experiment with that. However, I'm guessing that if I do that, then until I get a GUI for the non-Python folks, the only people that will help Puffin evolve are the Pythonistas. Now don't get me wrong, these are clearly the "chosen people" and Puffin could do a lot worse with theirs being the governing voices. But there are not all that many of them. If I leave it open (i.e. predominantly XML-based) then I get LOTS more eyes on the thing and make it better for EVERYONE. I will promise to experiment with some of your suggestions, though. And anything I pull together, I will make available on the puffin website. In the meantime, it's all in source forge. If anyone feels a hankerin' and has more free time than I do, have at it, but share... ;-) As for using WebUnit. I certainly do NOT have anything against it. It makes a lot of sense. I didn't use it the first time around because I didn't investigate it as much as I should have before too much water passed under the bridge. However, I'm revisiting several aspects of the fundamental architecture of Puffin to allow for such things as Email-based test actions, so I am in as good a position to change as ever I will be. I'll take another pass through the WebUnit docs, etc and see what it would take to replace what I have. Especially, if it were to give me stuff like NT Challenge/Response support, I may be VERY interested as this is a complete pain I've not incorporated yet. Keyton > -----Original Message----- > From: Chuck Esterbrook [mailto:[EMAIL PROTECTED]] > Sent: Thursday, June 06, 2002 9:17 PM > To: A. Keyton Weissinger; [EMAIL PROTECTED] > Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED] > Subject: Re: Web testing > > > On Thursday 06 June 2002 06:10 pm, A. Keyton Weissinger wrote: > > In the mean time, let me ask you (all) to consider this question: > > Would you still want a script-based version as much if a GUI > > prevented your exposure to the XML nastiness? In your experience with > > QA professionals, do you think it would be useful to QA folks? > > I can't answer the last question. > > Yes, I'd still want the Pythonic version over the GUI because of the > power that I get by having access to libraries. I mean, how else do I > iterate over the test users in the test database? And even providing > that, I will eventually come up with something. > > In any case, I don't see the conflict. Wouldn't this approach work?: > > - Make testing framework oriented towards Python, like I described. > - Make DeclarativePlan(Plan) class that can handle the XML declarative > format (and perhaps a Python dictionary declarative format) like what > you have now. > - Make a GUI for the convenience of declaring DeclarativePlan instances. > > > Now everyone can have cake and frosting! > > Also, consider using Purcell's framework (unless you have some beef > with it) as the bottom layer in this rich testing framework. > > Food for thought. I could be missing some issues. > > > -Chuck > -- > http://webware.sf.net/ > > _______________________________________________________________ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _______________________________________________ Webware-discuss mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/webware-discuss
