I was thinking that your group is going to be attending a conference...
that there are plenty of linguistic pundits who will try to
find a readability weakness or other chink in your armor.
Myself, I would avoid all balanced chars,
particularly < > . This would encourage
users to use yaml to wrap sgml,xml and html without quoting.
Particularly if & and other entities
cant occur where a yaml anchor is expected.
[or if they do, they are recognized as such]
m1htmlsnip: &id001 !htmlsnip <body><p>more stuff & even more</p></body>
m2htmlsnip: &id002 !htmlsnip <head><title>page assembled from
what you have now can already do the wrapping I think.
but simple minded editors (I will call them programmers-editors
for lack of a better euphemism) that count balanced parens and braces
probably start to flag things as "odd" when they see multiple
instances of yaml folded blocks.
when a newbie gets prodded by their editor to balance a document
that they know is already balanced, they are going to lose enthusiasm
and start to question your success at bringing out readability.
I won't cause I don't use such a weak editor but I am
also not in control of what others download.
"perhaps another character" -- replacements
I am thinking of alternatives for you and rejecting most of them.
your group made good picks from the punctuation set so far.
are the parsers you folks have made expecting a limited set of
punctuation and alpha at that point in a production?
| literal block on subsequent lines
> folded block on subsequent lines
" quoted string start (same with "'")
& anchor [*] (if alpha follows)
* refer-to-anchor [*] (if alpha follows)
! typespec [*] (if alpha follows)
newline and indent (go to sub blocks with defaults)
newline_no_additional_indent (more of same or illegal if no -, //, ',',
(nan|inf) and other numeric symbolizers [all have () ]
yes|no|true|false and other predefined scalar
alpha_upto_colon: url start ftp://site...blah mailto:me@...
alpha-no-colon string or number start
it looks to me that this is a parser reset/restart
point and possibly the only place
where [*] items can be introduced. I am doing this from memory
so I may have missed a character.
it also looks to me like ':' is illegal at that point currently.
perhaps same with // .
are other chars unexpected there?
At 02:34 PM 7/10/2002 -0400, you wrote:
>Sounds like you have some ideas... let us hear what
>you are thinking. We no longer use ] to indicate
>folded block scalar, but we use > instead. We don't
>have any other uses for the less than symbol, "<".
>(and XML uses the less than symbol to start its tag).
>Overall, I've found > to be rather clean looking... but
>this is probably subjective opinion; perhaps another character
>would serve us better?
>On Wed, Jul 10, 2002 at 08:50:55AM -0600, Bob Racko wrote:
>| I see that you have selected to use
>| a few of the "balanced" characters
>| (notably > and ] ) to indicate
>| special conditions in the parse sequence.
>| I understand that you find them to be necessary.
>| I would like to know where in the archives
>| I would find the prior discussion about those choices.
>| ;# mailto: bobr@...
>| This sf.net email is sponsored by:ThinkGeek
>| Two, two, TWO treats in one.
>| Yaml-core mailing list
>Clark C. Evans Axista, Inc.
>XCOLLA Collaborative Project Management Software
;# mailto: bobr@...