Python Progress Report Session

Panel Discussion on "What's Next"

Moderator: Barry Warsaw

• Panelists stories:
  • Paul: uses python for development on a contract basis
  • Mark: uses python purely for windows - high volume not really interesting in anything that's not native l&f
  • Jim: started using python to interface to lagacy code to windows. A way of using more modern data storage
  • Ted: CNRI's interest in python. Knowbot project (code mobility) and we just love the language
  • Mike: pushing the STEP? standard which is a product exchange standard. let's try using a tool as the mechanism for standard interfacing. (e.g. a printed circuit board as a python constructor)
    • build in usage protection and royalty collection into download protocol
    • PSA formed out of the LEOMA (Laser and Electro-optics Manf. Association) model
    • PSA as a quality control monitor <
• Let's turn it around: "What can we do?"
• Question and Answer
  • Q: what's being done on the convergence of DLL's for the various python ports for windows?
  • A: for the python core it's 99% there
  • Q: Do you develop on NT or 95?
  • A: NT and compile 95 binaries
  • Mark: the native look and feel issue is the most important thing to gain acceptance in the windows market
  • Jim: There is no good 3.11 port for python win32s is slow even with alot of memory, so a native 16bit app is the only *real* way to go, then you have all the addressing problems
  • The DLL for core python *may* get done if slight differences in console input/output can be addressed
  • Q: From a language issue, is there anything missing?
  • Paul: the language should not be a topic for the panel
  • Barry: A nice addition to the core library would be ILU or SYLU
  • Assertion: if we had a native l&f module that worked on every platform, we'd be better than *any* other language
  • Ken: last workshop's GUI session agreed that a base-line GUI (runs everywhere) in addition to the full blown platform dependant issues
  • Q: what are perceived impediments
  • Assertion: a common Object class like objective-C with methods like super for all objects
  • Response super can't be done with MI
  • Licencing Issues: Q: Is the fact that Grail downloads source a hinderince to commercial ventures?
  • Mark: Yes but so is cost (why Java is both good and bad) good that it obfuscates the source, bad cause it costs I'd like grail to grok pyc files, cause that's how I'd like to distribute software
  • SGI has taken python and changed it's name to momba? but that have contributed there plans back to the PSA but did not give credit to python or Guido... but (on the other hand) it was a very small research
  • Infoseek uses python, has enhance python performance, and gives alot of publicity to python
  • Proposal: GNU copyleft for Grail? CWI copyright for Python?
  • There have been quite a few companys that have been successful basing their products on gnu
  • Ted: Grail licencing. We did not want the likes of Microsoft or some other big organization with deep pockets to take over grail. At the same time, Bob is sensitive to commercial requirment for use of Grail
  • Proposal: require contributed software to have a certain copyright structure
  • Paul: what can the PSA do better?, documenting things?
  • Jim Fulton: I think we should have more contact with USENIX
  • On python.org having a very visible todo list to facilitate more voluteer work
  • reorganize the ftp sight (e.g contrib directory)
  • Ted: Bob Kahn wants to "pull the stops out" and really grow the PSA
  • Problem: the matrix folks are getting code from the matrix sight not python.org
  • Q: would you like my contributed NEXTSTEP Quad-fat binary for python? Does that screw with the pyc files?
  • Yes to 1) and No to 2)
  • Paul: We are doing more in the PSA and this third workshop went quite well!

Python Workshop Feedback Session