[llvm-dev] RFC: changing variable naming rules in LLVM codebase

Chandler Carruth via llvm-dev llvm-dev at lists.llvm.org
Mon Feb 18 18:08:22 PST 2019

On Mon, Feb 18, 2019 at 10:10 AM Krzysztof Parzyszek via llvm-dev <
llvm-dev at lists.llvm.org> wrote:

> On 2/18/2019 4:15 AM, Michael Platings via llvm-dev wrote:
> > Taking my previous example [1]:
> >
> > InnerLoopVectorizer LB(L, PSE, LI, DT, TLI, TTI, AC, ORE, VF.Width, IC,
> >    &LVL, &CM);
> >
> > If we imagine that over time it evolves such that 50% of the variables
> have been renamed to camelBack versions of the type names, then it will
> look like this:
> >
> > InnerLoopVectorizer LB(loop, PSE, loopInfo, DT, targetLibraryInfo, TTI,
> >    assumptionCache, ORE, vectorizationFactor.Width, IC,
> >    &loopVectorizationLegality, &CM);
> Hold on...
> The change from UpperCamel to lowerCamel should be separate from going
> from X to somethingOtherName.

FWIW, I suspect separating the transition of our acronyms from the
transition of identifiers with non-acronym words may be an effective way to
chip away at the transition cost... Definitely an area that people who
really care about this should look at carefully.

> It seems like in this example, TLI is changed to targetLibraryInfo for
> the purpose of having a name that lowerCamel can be applied to, which
> is, at best, backwards.
> When a new person sees "TLI", they won't know what it is. When an LLVM
> developer sees "TLI" they know exactly what it is, even without any
> context. At the same time, a person is new to LLVM for only a certain
> period of time, much shorter than the lifetime of the code.
> The key to readability is to make the important things easy to see, and
> get the unimportant things out of the way. By using completely expanded
> names we run the risk of making everything equally "easy to see"...

I think this bias towards acronyms (which I used to share) due to keeping
things short but still recognizable once people become deeply familiar with
LLVM is the wrong prioritization. It does work well for experienced LLVM
developers, but I think we should do much more to facilitate and encourage
people who are not in this set. While this does come at some cost to highly
experienced LLVM developers (reading `library_info` instead of `TLI`), but
it seems easily worth it to make the codebase more accessible to new
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20190218/8d82b5ef/attachment.html>

More information about the llvm-dev mailing list