Re: [gopher] Which itemtype for video (vote)?

2012/5/20 Kacper Gutowski <mwgamera@gmail.com>

On 2012-05-19 16:56:28, Jeff W wrote:
> Nick Matavka <n.theodore.matavka.files@gmail.com> wrote:
> > Gentlemen and ladies:
> > It has come to my attention that the choice of item-type for video is
> > somewhat controversial. The existing, but somewhat awkward, choice is
> > ';' (semicolon), which is my own preference, but there are some who
> > have floated the idea of it being changed to 'v' (lowercase vee). The
> > sole reason I prefer the semicolon is that it is the (unofficially)
> > established standard. If the selector were changed, admins would have
> > to update their sites to accomplish the change, and some Gopher sites
> > are... shall we say... updated once in a bit over a millennium.
> > Furthermore, clients would have to be reprogrammed to accept the 'v'
> > selector as video.
> >
> > vee: 0
> > semicolon: 1
>
> Simple votes:
>
> "v" for video: 1
> vague control character-looking semi-colon: 0
>
> Alternately, I'm fine with lumping all the binaries under "9"
>
> Civic duty - done.

I'd prefer using '9' for video because, well, these are binary files.
If you need to know whether it's a video of something else you can guess
it either by selector (extension) as RFC 1436 suggests or using some
magic numbers. If you want to play it, you'll have to do it anyway.

RFC 1436 states that "Characters '0' through 'Z' are reserved. Local
experiments should use other characters." Character ';' happens to fall
in the '0'-'Z' range (at least in ASCII) which was probably an error in
the first place. But while I'm not sure what body is in power to make
use of reserved characters, I think that once consensus is reached that
the type is indeed needed, it should get "standardised" in some way and
get one of the characters that were reserved for this purpose.

If we really need separate type for video and we want to standardise it,
I'd highly recommend using already common and widespread ';' because
it's already in the reserved range so client getting ';' can not even
safely assume it's some invalid experimental stuff, it's a type it really
should know about.

For now, we have a lot of guesswork on client side. I think that the
most common and widespread non-standard types like 'h', 'i', or ';'
should get adopted as de-facto standard. But adding another character
like 'v' to the mess won't make things any better.

The fact that ';' is "vague control-character-looking" is a weak
argument as item type was never meant to be shown to end-users and
numbers for standard types aren't more informative anyway.

semicolon: 2
vee: 1
(don't bother and use nine: 2)

K.

_______________________________________________
Gopher-Project mailing list
Gopher-Project@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/gopher-project