- From: Graydon <graydonish@gmail.com>
- Date: Wed, 29 Jan 2025 10:36:47 -0500
- To: Bethan Tovey-Walsh <bytheway@linguacelta.com>
- Cc: ixml <public-ixml@w3.org>
On Tue, Jan 28, 2025 at 05:50:50PM +0000, Bethan Tovey-Walsh scripsit: > Seconding David's thanks for this useful distinction, Graydon. (Also, > off-topic, but I find "fails to delight" paradoxically delightful.) I cannot but feel pleased that this phrasing has proved effective! (You're both welcome. I have been lucky enough to use iXML in production so I might have more clarity on the subject just from experience.) > It'll be interesting to see how introducing a negation operator (if we > manage to do so) expands the class of problems for which iXML is a > good solution. I am interested in a negation operator, I think it's potentially useful, but a substantial constraint on "good solution" is "Can I debug this when I come back to it after a year?" and (data point of one!) iXML grammars as they already exist are horrid things to debug even after one has spent the time to stuff the whole majestic expanse of the specific grammar back into one's brain. Debugging tools would to my mind be the most useful things for driving iXML adoption. A processor switch to emit "this boils down to the following" simplified versions of what one wrote where the switch knows it's trying to help a would-be grammarian would be the first thing on my personal list, but nigh-anything in the way of debugging help would improve matters. (By analogy to ancient days, a linter might be the way to go. I cannot claim sufficient comprehension of the issues around writing such a thing to have an informed opinion, but I do know that debugging iXML grammars feels a lot like "segmentation fault, core dumped".) So, anyway; I think iXML is already "who will use this?" limited. I suspect that group could be expanded through tool support much more than through expressiveness of the grammars. -- Graydon -- Graydon Saunders | graydonish@fastmail.com Þæs oferéode, ðisses swá mæg. -- Deor ("That passed, so may this.")
Received on Wednesday, 29 January 2025 15:36:53 UTC