[bind10-dev] observation

[Evan Hunt]

> I don't have a strong opinion about RRClass and RRType, although I
> personally wouldn't necessarily consider RRType/RRClass vs RRset a
> style inconsistency.  If others strongly prefer the consistent use of
> the case for the RR* classes, I'm willing to change that.  In that
> case, since the style of RRset has (IMO) a compelling reason, we'd
> change RRClass/RRType to RRclass/RRtype respectively.

Based on the errors I've made so far, I misspelled RRset (by using a
capital S) much less often than I misspelled RRClass and RRType (by
using lower-case c and t).  Something like a tenth as often.  So
standardizing on RRx would apparently match my own fingers' bias, in
addition to your other good reasons.

(This is, btw, one of the reasons I complained about CamelCase in the
first place: it's not just the extra keystrokes, it's the fact that
things like RDataSet and RdataSet and RDataset are comparatively difficult
to visually distinguish and to remember, and this can lead to errors.
However, I don't want to reopen that debate now!)

If anyone else is finding it difficult to remember where the capitals go,
it might be worth considering a switch to RRclass and RRtype, is all I'm
saying.  If it's just me, I'll adapt.

> One consideration point would be the RRTTL class.  Following this
> convention, it would become RRttl, but this seems to be a bit awkward
> to me, in that lower case letters are used for an acronym
> (i.e. "ttl").

I think RRTTL is a reasonable place to make an exception on this.

                                        eh