This seems reminiscent of a podcast I heard about Elephant 2000: http://www-formal.stanford.edu/jmc/elephant/elephant.html http://feeds.conversationsnetwork.org/~r/channel/itc/~3/456508493/detail3770.html
His goal with that language is to create a system which uses English to define the interaction, but the compiler develops the details. (or this may be completely off target....) On Wed, Dec 3, 2008 at 4:30 PM, Edward K. Ream <[EMAIL PROTECTED]> wrote: > > On Wed, Dec 3, 2008 at 3:00 PM, ne1uno <[EMAIL PROTECTED]> wrote: > > > a few rambling thoughts. it is always about the editor! > > The fun thing about the creative phase is that contradiction works. > You can say, "it's about the editor" and I'll so, yes, in a sense. > And yet the project is more exciting than doing yet another gui for > Leo :-) > > > sounds to me like a deeper study of AWK is in order. > > The pattern matcher? Pattern matching may be a big part. > > > I have not needed pythoscope or lib2to3. is it for python3k? > > Developed for py3k, but it has more general uses. > > > there is a huge > > need for easier data visualization to enable more flexible > > translation tasks. > > I agree. I still think of Sherlock fondly. Adding Sherlock features > to Python's Logger would help it a lot. > > Anyway, clever visualization is indeed very important, both as an end > product and as a way of studying complex code. > > > maybe data translation is program generation afterall. > > I may not be talking about program generation as people typically do. > Everything is blurry now, which is a good sign. > > > more scriptability, easier to grasp API's, self healing, > > are they enabled with better completion and help lookup? > > some new combination of libs? we seem to love to invalidate > > most of the work already accomplished every few (months,) years > > in order to arrive at the need for ever more tricky translation > > tasks just to break even. prodigious amounts of hardware and > > software upgrades littering the path to there > > as well as lost efficiencies of scale when the installed base shrinks. > > Well, nobody wants to hold on to the Wright Brothers first flyer :-) > > > more concretely, how great would it be to further enhanced the > > ability of plugins to troubleshoot themselves instead of the current > > situation of waiting for new or renewed users to provide traceback > > after a failure. > > I want pythoscope to generate test cases for plugins. That would be > better than the present nothing. > > > has anyone succeeded in teaching a computer the intent of a plugin? > > so we make better debuggers for more trials and little progress. > > I like your discontent :-) It's the spur to improvements. I'm not > sure we are discontented about the same things, but probably that > doesn't matter. > > > can I suggest a first order of business is to divorce Leo > > from python? as good as it has been for Leo it is now a drag. > > There is nothing half as good as Python. If you need C++, the only > proper thing to do is to wrap it in Python. Usually that isn't > needed, but there is no way I'm ever going to mess with C++ again. > That would be a big step backwards. > > > as comfortable as python may well be. several orders of > > newer versions will still leave us basically where we are. > > Why do you say that? What we need are higher and higher tools. We > can't begin to get those tools messing with C++. > > > we haven't begun to mine the > > archives for data visualization and translation technologies > > that have at various times been brought up in connection with > > Leo as the bridge if only we could see all the pieces at once > > and be there with a grasp of how to get from here to there. > > Leo's breakthroughs have been mostly ideas: script buttons, @auto, > @shadow. ILeo and the minibuffer may be exceptions: the > implementation was (mostly) everything. > > > it often takes someone on both sides to meet in the middle. > > Many people here can contribute on both sides. > > > so what can pythoscope do that AWK hasn't > > already mapped out as possible 40 years ago? > > I think you are bit pessimistic :-) As you say, details matter. I > know for sure that Python + Tkinter opened doors that were never > visible with Borland C++, though theoretically both platforms were > equivalent. > > Rather than worrying about what is possible, the thing to do now is to > read widely, looking for a) what works and b) what doesn't work. Both > are valuable, and both contribute to progress. > > Edward > > > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "leo-editor" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/leo-editor?hl=en -~----------~----~----~----~------~----~------~--~---
