On Tue, 2010-02-23 at 17:45 -0800, Brian Swetland wrote:
> On Tue, Feb 23, 2010 at 4:08 PM, Daniel Walker <dwalker@xxxxxxxxxxxxxx> wrote:
> > I wanted to have the fish names discussion one more time. I think it's
> > fine for Google to use those names internally, but we really want actual
> > product names for the permanent kernel history.
> >
> > I know Pavel was the first to bring this up, and I do think it makes
> > sense. I don't know if board-dream.c is the exact right name to have,
> > but I think it's more along the right lines.
> >
> > I also think it makes sense to register real names for the
> > machine_is_xxx names .. Since when your writing code it's easier to know
> > what machine_is_nexusone() is vs. machine_is_mahimahi() ..
> >
> > I'm not suggesting you guys not use fish names altogether, I'm just
> > concerned how this naming will impact the programmers outside Google who
> > will be touching this code, and the general user population.
>
> A big issue with the fish names is that final product names do not
> always happen until late in the process (Nexus One was very far along
> by the time it ended up being called Nexus One, for example). It is
> not at all uncommon for hardware products to have names that they're
> known by which are not the mass market names, and also not uncommon
> for these products to have a single board (trout, for example) which
> is instantiated in a number of mass market products (Dream, G1, ADP1,
> Revolution, etc).
>
> The bootloaders for these products identify themselves using the
> board/machine names, the userspace code sometimes locates features
> based on those names, etc.
>
> I still fail to understand why there's a problem with naming the
> boards based on the names the development team for the device used.
> That these names happen to be fish instead of a collection of numbers
> and letters or some other codename seems unimportant.
>
> This really feels like an needless hurdle ("we don't like your board
> names") rather than a valid issue ("code does not compile", "code
> fails checkpatch / has style violations", etc).
It really is important tho. It's because the vast majority of people who
will work on this code will not work for Google, and may not easily know
which fish goes with which product .. We're talking about people
actually writing code, or fixing defect etc.
Just look at Dream for instance, you guys don't do dream work anymore.
However, the phones are still out there, with tons of Dev phones. I'm
still doing work on Dream right now for mainline .. So you have to count
up all the times people spend scratching their heads year after year
saying "Which fish is this?" It makes tons more sense for it to just be
explicit.
Daniel
--
To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
