Another thought on the difference between the two methods is that ll_get_attribute(license_uri, LL_ATTRIBUTE_URI_CREATOR, 0)
would allow third-party attributes to be used, without altering the liblicense API. However, I'm not sure it that's something we want to encourage or not. I also think that ll_get_attribute should leave off the last parameter and we'd include another method, say, ll_get_attribute_localized. The most common use of ll_get_attribute doesn't use the last parameter, so the API would be cleaner for the most common case. Cheers, Jason > Today on #cc there was some discussion about the use of helper functions > vs the use of named constant URIs. This came about as a comment by > paulproteus (sorry, don't know mapping from IRC handle to email address) > on a recent commit by me (pmiller) of the ll_license_get_creator and > ll_license_get_publisher functions. > > The question was: which of the following is better in the public API? > > ll_get_attribute(license_uri, LL_ATTRIBUTE_URI_CREATOR, 0) > > or > > ll_license_get_creator(license_uri) > > The second form is quite simply a wrapper function which uses the first > form. > > My feeling is that both are desirable. The first is great for > generality and experts, when code wants to loop over every attribute in > a license. The second form simply has less for beginners to get wrong. > Both forms can be made equally robust in the face of changes in the > spelling of the actual attribute URIs. We already had wrapper functions > like this before ll_license_get_creator and ll_license_get_publisher > were committed, so the API remains self consistent. > > We agree on something: neither of us likes the > "http://creativecommons.org/ns#blah"; constant strings in the code. They > should appear exactly once, and replace uses with named constants. > > > A second point of discussion: which of the following forms should be > used to define the constants? > > #define LL_ATTRIBUTE_URI_CREATOR "http://blah..."; > > or > > extern const char LL_ATTRIBUTE_URI_CREATOR[]; > > we both like the second form, but for different reasons. Robustness > mentioned above would use the second form. > > > Attached is a patch which implements what we were talking about, or at > least what I *think* we were talking about. I haven't committed it yet, > I would like some discussion first. > > > Regards > Peter Miller <[EMAIL PROTECTED]> > /\/\* http://miller.emu.id.au/pmiller/ > > PGP public key ID: 1024D/D0EDB64D > fingerprint = AD0A C5DF C426 4F03 5D53 2BDB 18D8 A4E2 D0ED B64D > See http://www.keyserver.net or any PGP keyserver for public key. > > "This is terribly unsatisfying." "It's far more satisfying if you throw > it at someone." -- Final Cut Pro easter egg > > ------------------------------------------------------------------------ > > _______________________________________________ > cc-devel mailing list > [email protected] > http://lists.ibiblio.org/mailman/listinfo/cc-devel _______________________________________________ cc-devel mailing list [email protected] http://lists.ibiblio.org/mailman/listinfo/cc-devel
