Hacker News Reader

Comment by fmax30 | Hacker News Reader

Mozilla MDN docs are my go-to resource for js related stuff, does this mean they won't be maintained and updated anymore?

To be honest, I would be more than willing to pay a reasonable monthly subscription fee to access their docs.

View on HN

It’s not only them anymore: "Mozilla brings Microsoft, Google, the W3C, Samsung together to create cross-browser documentation on MDN"

-- https://blog.mozilla.org/blog/2017/10/18/mozilla-brings-micr...

That’s a blog post from 2017. I wouldn’t be so sure it will continue under the Mozilla umbrella without them dedicating any people to it.

That's at least a tiny bit of good news then. Thanks for clarifying that!

The blog is from 2017!

That is good news. Do any of the other companies have people working on MDN?

I agree, these are the best JS docs around. I hope they continue maintaining them. The main alternative is W3 Schools, which is substantially worse (I blame their for-profit ad-supported business model, which doesn't necessarily incentivize thoroughness and depth)

I am constantly surprised to see Google ranking W3 Schools so highly. Does anyone know how they have managed to be ranked so high up?

It’s decent for beginners with easy to follow tutorials and examples, but once you start getting in the weeds of JS, it’s not as useful I feel like.

Domain Age - They've been around forever with tons of materials.

Quality - good enough for most things. The newer stuff is not that bad and usually more practical(read quicker to grok for someone in a hurry) that the more thorough MDN docs.

Dirty trick - association with w3 standards body (none in reality)

They used to be really bad. Their quality has significantly increased over the past years. I will use w3schools for quick info, but MDN when I really want in-depth information.

https://www.ecma-international.org/ecma-262/ is much more comprehensive, and sadly very few ever read it.

Lots of functionality in the spec that isn't even possible to find on MDN or W3 Schools.

Not sure whether it's fair to compare them. Specifications are usually written using a format language for the ones who implement them and not designed to be a learning material, so many things which are not necessary for this goal are omitted. On the other hand references like MDN are dedicated to the end user, written with a regular language and unnecessary details removed but also with additional explanations and usage examples which are really helpful. So one usually doesn't need to refer to the specification unless there are reasons to do it(e.g. need to understand all the aspects of a feature). It's pretty hard to understand what's written there for a beginner without some context which provided by manuals like MDN.

Be sure to read https://tc39.es/ecma262/ instead of the one hosted on ecma-international.org that has up to a year's worth of unfixed bugs.

A while ago they established an advisory board with members from institutions with deep pockets so if Mozilla were to abandon MDN entirely, chances are we might see a more democratized version emerge, which is a good thing. I'd wait till the board meets again this month. But hey if that didn't happen, MDN content so far is fairly open - CC-BY-SA, public domain and MIT, so some kinda fork is bound to popup.

MDN is a Wiki. Everybody can contribute.

Doesn't mean everybody will. Wikipedia is a weird special-case. I've contributed once to MDN, ever, and I use it all the time.

Also if the team that was supporting it is gone, will the deployment just bit rot and die?

I'm not even sure I want everyone contributing, at least not without some pruning.

Sparse documentation is better than wrong documentation.

I've never come across a page on MDN I felt I could contribute to since it's so complete, but if I were to come across out of date docs I absolutely would. I imagine it's similar for a lot of folks.

MDN has some Python tutorials that are of lesser quality than HTML,CSS,JS MDN docs.

Those could use some work.

Usually there are only a handful percents of overall users who actually create content compared to "read-only" users. There is nothing special about wikipedia, the same thing applies for sites like youtube,reddit, etc. The same will apply for MDN as well, so if you haven't contributed there, this doesn't mean that nobody will.

>Wikipedia is a weird special-case.

Is it?

https://www.vice.com/en_us/article/7x47bb/wikipedia-editors-...

This is the nature of any community platform [0]. Most people will just fix small issues they see or contribute a little in an area that appeals to them, and aren't interested in becoming heavily active. I don't see Wikipedia as special in this case, it's just so large that the 1% has enough critical mass to maintain it (well, mostly).

0: https://en.wikipedia.org/wiki/1%25_rule_(Internet_culture)

>Most people will just fix small issues they see or contribute a little in an area that appeals to them

Wikipedia makes this very difficult as well. I had to submit a small update to sales numbers for the best-selling video game franchises of all time THREE times because two separate admins reverted my edits for the most stupid reasons possible. Then they had some hand-wringing about the lack of sources (even though most of the data there wasn't sourced any better) before they stubbornly decided to stop reverting my edits. A less stubborn person would've given up editing Wikipedia after the first or even the second reversion.

But e.g. Google teams working on web specs do at least. But I don't have data how the contributions are split, and maintenance is of course a concern.

Is the source data openly available, though? You can't exactly "fork" a wiki, normally. I'm concerned lots of crucial knowledge would be lost if Mozilla decided not to host it any more.

Many wikis, MDN included, have all content under an open license, and even run open-source software, and thus can be forked if necessary.

But is the content made readily available in the original markup (wiki markup, markdown, whatever), or would you have to scrape everything from the website? Wiki content is normally stored in a database, and presumably that database isn't just open to the world to query directly

Many wikis have open APIs (e.g. mediawiki instances typically do). Not sure about MDNs custom thing, but scraping is always an alternative - or even working off one of the HTML dumps available for offline use (although loosing history and original format sucks).

Theoretically since you can edit the original-format content from the website, you could scrape it from the website. I'm just wondering if this is something anybody is thinking about/working on.

Same here, W3 pales in comparison, which is not to denigrate their work. It's been helpful many times.

MDN just has more accessible writing and notes on best practice.

That's not an accident, IMO.

W3 docs started out, and I would argue, continues to be most relevant to people whose work it is to implement a browser.

Certainly, you can use them to learn how to use the various HTML, CSS, and JS standards, but this is where MDN - and even the Chromium and Webkit resources - are leaps ahead with practical examples you can use in your own code.

I think people are talking about w3schools, where you're talking about the specs that the W3 (in some cases) maintain.

Really, I wasn't sure when I wrote that. Refsnes Data, who owns/runs the site, isn't affiliated with the W3C in any official capacity that I know of.

I was thinking more of a story I heard on an old episode of Hypercritical where John described first learning HTML and the rest through a combination of (mostly) O'Reilly books and reading the specs themselves.

While I suppose this passed muster in the early/mid 90's, I would repeat my earlier claim that the specs nowadays are more for implementers, and not something really well suited to teach you how to use the various languages and tools.

Potentially losing what is perhaps the most approachable reference for this purpose would be terrible for everyone.

> I would repeat my earlier claim that the specs nowadays are more for implementers

Right... but that's not what anyone here means when they talk about the w3 compared to MDN. They mean w3schools.

Hope someone scrapes them and hosts them as a torrent now, I'm not sure where else to go for js documentation besides node.

yeah, i google my questions about js/hmtl/css and always look for MDN docs. would be a huge loss to lose that

Reply on news.ycombinator.com