[osis-core] Lists in Attribute values: final call

[Steven J. DeRose]

At 18:29 -0600 2003-10-20, Todd Tillinghast wrote:
>Chris,
>
>Assuming xA0 is not whitespace as Patrick suggests, this would be a good
>way to keep spaces and hyphens separate when substituted for xA0 and _
>respectively).

As in many systems, xA0 is not considered a whitespace character -- 
which is why software will naturally not break at it.

On the other hand, it's also not an XML name character, so not 
permitted in attributes of types like MLTOKENS at all.

I think we should stick with space-delimited values for these reasons:

1) We're in XML, and that's what XML does. I think the chance of that 
changing anytime is very low, because:

    a) it won't become worth it to break backward compatibility

    b) No one will be able to agree on what delimiter to use instead

2) We already use space-delimitation, such as in osisRef, so by 
sticking with it we stay simpler

3) Yes, there are reference and naming systems out there that use 
spaces. Like, say, "1 Tim 2:2". Oh wait, we already force people to 
not use spaces (or colons) that way. In other words, I'm saying that 
the major case of space usage, we've already decided not to go along 
with; I think the remaining cases are few enough that saying either 
"substitute "_" or delete the space (when feasible) is ok.

4) We *have* to reserve some delimiters for ourselves -- and we've 
already reserved space, colon, period, and hyphen; so I think we 
should just stick with reserving that list (brackets and @ are only 
reserved in a certain context).

5) "|" is prohibited in URLs, which is not a huge deal (so is space, 
though not colon, dot, and @), but would add to inconvenience from 
time to time (esp. since I think we should define our reference 
system as a full-fledged URN scheme sometimes fairly soon).

S


-- 

Steve DeRose -- http://www.derose.net
Chair, Bible Technologies Group -- http://www.bibletechnologies.net
Email: sderose@acm.org  or  steve@derose.net