[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [Openstack-devel] Bug#708658: Bug#708658: [RFR] templates://cinder/{cinder-common.templates, cinder-api.templates}

(Leaving the buglog off Cc, as this certainly isn't a bug in Cinder)

Thomas Goirand wrote:
> Justin B Rye wrote:
>> No, it's far too late to change it!  But presumably it's "creeping
>> developerese": programmers forget that what API actually means is
>> "design specs for source code" and think of it as meaning just "the
>> part of the software I interact with".  Even "ABI" would be closer -
>> at least that's a runtime thing!
> Not that much though. OpenStack also has python-<whatever>client which
> really does contain Application Programers Interface in python.

Yes, the package python-cinderclient provides an API to allow Python
programmers to write cinder clients, but it's equally true that
python-aalib provides an API to allow Python programmers to handle
aalib, and python-zzuf provides an API to allow Python programmers to
handle zzuf.  This isn't evidence that aalib, cinderclient, or zzuf
are themselves APIs.

Originally, APIs were paper spec documents for pieces of software;
object oriented languages let you talk more abstractly about a set of 
public methods as forming an API, so eventually programmers have come
to think of even things like the protocol used for communicating with
a web service as an "API".  But it worries me a little that they don't
seem to be aware that this metaphor is getting rather stretched, and
that comparing something to an API won't make it any more readily
intelligible for the general public!
JBR	with qualifications in linguistics, experience as a Debian
	sysadmin, and probably no clue about this particular package

Reply to: