On Thu, 11 May 2000, you wrote: # -----BEGIN PGP SIGNED MESSAGE----- # Hash: SHA1 # # > # > Ick and double ick. The data type and interpretation of field # > values is already 100% completely specified by NAME. If you don't # > already know the type of a field by its name, you shouldn't be # > reading it. This just generates another source of errors: now I # > have to write code that not only grabs a field by name, but now I # > have to check whether the type matches what I expect. # > # > If you think you need data types so that a code layer interposed # > between the code that reads the message and the code that uses # > the fields can do conversions, that's just a layer of code that # > doesn't need to be there. # # Lee, what you're forcing by saying "All data are strings" is that there is # no room to write any other protocol than the text one.. You can't have # optimized representations of integers, different string encoding formats, # You can't have newlines in strings, etc. #
If you really need to embed non-printables into strings, use the \code value. I happen to be a python guy, but I find repr() to be very useful in this context. After all, 1, 0x01, and 01 all mean the same thing eventually. -- Ian Zepp [devel at edotorg.org] KDE & Python developer [http://www.edotorg.org] _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
