W3C home > Mailing lists > Public > www-validator@w3.org > November 2017

RE: Building blocks using data

From: <jamectomy@tutanota.com>
Date: Mon, 27 Nov 2017 00:56:10 +0100 (CET)
To: Www Validator <www-validator@w3.org>
Message-ID: <Kzt48J_--3-0@tutanota.com>
I didn't mean to trivialize what you do and I apologize for wording it that way. If this post makes it hard for you to respond respectfully please pass it by. Constructive and thoughtful input is welcome. I'll try to restate my question for other people that might be interested in helping.

I'm not trying to make a WYSIWYG editor. It's almost the opposite. I'm trying to make something like https://developers.google.com/blockly for editing html code. Take a quick look at the link before making suggestions because I think it'd really clarify a lot that's hard to put into words. It wouldn't look like those colored jigsaw blocks it would just work in a similar way. It would look more like regular html code looks.

Thank you for the links to documentation. I'm not currently looking for documentation. I'm looking for a file or files that I can use to parse what elements are allowed in what other elements and so on. From the searching I've done it looks like what would be most useful is some kind of validator codebase that contains data describing the relationships between elements and so on, but the closest document I've found to that has been in the validator files for html4 at https://dvcs.w3.org/hg/markup-validator/file/tip/htdocs/sgml-lib/ISO-HTML/15445.dtd. I'm looking for the same general type of thing for the more straightforward parts of html5. This doesn't have to be the complex stuff just things like the fact that divs can't go directly into table elements. A simplified JSON object of the same thing might look something like this (though this example is incomplete of course)

"head":{"canContain":["meta", "title","link"],"attributes":[...]},

I'm not sure what the previous response meant about loops and logic. Html doesn't have it, but html validators do. The link above shows how dtd is/was used to validate xml and xhtml. I read a few different things that looked like they said html5 is not validated that way anymore. I read that it's validated using schemas and run-time code. An example is a stackoverflow response at https://stackoverflow.com/a/15245834. It says "...HTML5 validators (basically the HTML5 mode of http://validator.nu and its  copy at http://validator.w3.org) use schemas and ad hoc checks...". Another is the about section of the validator site at https://about.validator.nu/#no-dtd. Maybe the schemas are what I'm looking for, but I haven't had any luck finding them.

I'm hoping to build off each other's work and keep figuring out more cool stuff to do with html. Part of the great thing about working with others is building on top of each other's work to make more progress. We all build on the shoulders of the giants that came before us. If nothing like this exists that's ok too. I just thought a validator might have the relationship data laid out like that somewhere in its code, but I haven't been able to find it.

Also it sounds like the irc in general is an abandoned technology. This might be its replacement but if anyone knows about a chatroom type place to talk about validator stuff let me know.

asl --> 76, blue, 'https://lists.w3.org/Archives/Public/www-validator/2017Nov/' + n + '.html'

From: O'Guin, Phaewryn D. <JDO09280@ccv.vsc.edu>
Date: Tue, 21 Nov 2017 09:22:42 +0000
To: "www-validator@w3.org" <www-validator@w3.org> 

HTML validation is a check after the code is written. Also, HTML doesn’t work like scripting does, so not sure what you’re trying to do here. There are no loops or conditional logic in HTML. I think what you’re trying to do has already been done, it’s called WYSIWYG and it allows newbs to drag and drop elements to make HTML documents/pages, and it’s terrible.

But https://www.w3.org/WebPlatform/WG/PubStatus#HTML_specifications may be helpful? Also https://www.w3schools.com/html/default.asp shows many attributes for a given tag.

Standards change, if you’re not willing to update your program to reflect those charges, the world is better off without your program. Development requires dedication. If you don’t plan to swim regularly for the rest of your life, there’s no point in buying a swimsuit. Splash around in the shallow end, but once you jump in the deep end, you can never leave.

Do people still use IRC? wow. I suspect you’re on the replacement. hi! asl? ;P


Phaewryn (J.D.) O’Guin

From: jamectomy@tutanota.com [mailto:jamectomy@tutanota.com]
Sent: Monday, November 20, 2017 7:20 PM - 19:20 PM
To: www-validator@w3.org
Subject: Building blocks using data

I'm trying to make code to make blockly-type blocks for html. This is blockly --> https://developers.google.com/blockly<https://developers.google.com/blockly>. This needs to be predictive not a check after code has already been written. I know html5 isn't completely serializable but some of it is. For example input elements have a type attribute. If I had a good source for scraping those very basic types of relationships between some of the elements and some other elements and attributes then it's possible I could parse/use that data to build my blocks.

Does anyone have advice about where I can find an up-to-date source of data? If possible I'd like data that works for Chromium but I'll take what I can get. Does any source exist in a way that could be useful to me or is all this validation done purely with logic now? One idea is to scrape the MDN site. I could scrape Chromium code possibly.

Does anyone have any other advice about how I could go about this other than manually writing out the data and then manually updating whenever anything changes?

p.s. the freenode channel is supposed to be #validator but there was nobody there when I went on. I saw it listed at https://validator.w3.org/feedback..html#mailinglist. Is it somewhere new now?
Received on Monday, 27 November 2017 00:30:07 UTC

This archive was generated by hypermail 2.3.1 : Monday, 27 November 2017 00:30:11 UTC