Reject away, but I'm afraid you've still got some work to do to
convince me that PEP 8 is more work for an SR system than any other
convention.
Name <cap>name
higher than normal recognition error rate. can require multiple tries
or hand
correction
MultiWordName <cap>mulit<no-space><cap>word<nospace<cap>name
very high error rate. many retries or hand hurting typing.
multi_word_name multi<underscore>word<underscore>name
normal error rate (low), can need multiple tries or hand correction
It sounds like the issue should be one of making your
screen-reader smarter, not dumbing down Python conventions. I
don't know what SR you're using (Jaws? Window Eyes? yasr?
screeder? speakup? VoiceOver?) but it sounds like at least for
the above cases, along with the PEP-8 MULTI_WORD_NAME constant
convention, a simple regexp+transformation should be able to
reprocess the input into something easier to
handle/hear/understand. I'm not sure any/all of the
previously-listed screen-readers give such regexp transformation
control, but I would expect that at least the OSS ones (yasr,
screeder, & speakup) it would be possible to add the feature if
it doesn't already exist. For these three, you might ping the
blinux mailing list to see if anybody there knows how to
implement such transforms.
StdlYCps <cap>sierra tango delta lima <cap>yankee <cap>charley papa sierra
*** very high error rate *** search and replace for all instances with
a x_y_z
form name is recommended
As for StuDlyCaps, it's hard on the seeing too, so I advocate a
firm smack on the back of the head for those that prefer this
abomination. :-)
-tkc
--
http://mail.python.org/mailman/listinfo/python-list
