On Tuesday, February 26, 2013 08:20:54 Jacob Carlborg wrote: > On 2013-02-25 19:21, Dmitry Olshansky wrote: > > What can I say Phobos is an example of software evolution ;) > > Is the new std.uni completely backwards compatible with the old one. > Otherwise we have the same problem as with std.process. But in this case > we have another name that is actually better. > > I'm having some problem understanding the obsession in shortening names, > even when the full name isn't particular long.
Well, it can get pretty bad with module names when you're forced to give the full import path. For instance, std.string, std.ascii, and std.uni all have toLower, and std.unicode.toLower is definitely longer than std.uni.toLower. In general, names should be as long as they need to be in order to be properly clear and descriptive but no longer. Making names too short makes code harder to read and understand, and making them too long makes it harder to fit as much in a line of code without it getting too long. That being said, std.unicode is probably a better name than std.uni, but at this point, it's better to maintain backwards compatibility than to rename it. If we need to rename it because of changes in the API, then going with std.unicode makes sense, but if the necessary changes are backwards compatible, then we should avoid renaming it and thus avoid breaking code. - Jonathan M Davis
