My background: In my free time, I teach web development to complete beginners, for the SF GirlDevelopIt chapter (using http://www.teaching-materials.org/).
A few observations I've made:
1) For the sort of things that they are trying to learn (basic HTML tags), W3Schools is a great resource. It's clear and fairly up to date. No other resource has managed to overtake them for beginner friendliness. I was hoping http://docs.webplatform.org/wiki/Main_Page might, but right now it seems like just a skin over MDN and searching doesn't work so well. I don't know if they've given up or what. (One idea: You could help them)
2) After I teach the second HTML lesson, I will explicitly show students "okay, here's how you search for HTML help on the internet." At that point, I'll show them W3Schools and MDN and explain the drawbacks/benefits of each. Don't know how much of that sticks with them, but it's worth a shot.
3) It took me a long time to put together HTML/CSS/JS curriculum. I first started several years ago, when teaching our first workshop, and have been adding to it since (and there are still more topics I want to add). I did also start video taping my lectures, as I figured it could be a resource for people who didn't make it to the workshops, and that also took me a long time - about an hour to record and edit a 10 minute video. So, I support anyone putting out more great knowledge material in the world, I just hope you have a fair amount of time allocated to work on this, so that you can it proper justice.
4) I recommend having a test audience for your material, like somebody that you tutor, so you can see how they respond to it and where their knowledge holes are. Also, a feedback button. :-)
In regard to (1), I've been finding recently that I'll look up a CSS property, and it lists the properties and has an example, but I find it's really crap for actually learning anything because the example doesn't actually show you the output. And more than example with screenshots and explanations of each would be useful to actually learn how to use it...
2) Juvenile language and problem statement. Off putting pitch to developers that found W3Schools useful. At some point it'll have to switch to a more professional approach.
3) Reading between the lines, apparently the author's goal is to crowdsource basic tutorial development with a long term for profit motive involving premium content. While transparent to mention this up front, it leaves me uninterested in participation.
Yeah i got that too cant people just get a book and read it - thats what I did on my first web project back in 94.
Played around at lunch times a BT and volunteered for a month in Edinburgh on secondment brought the only book on HTML published at the time and read it through a couple of times on the train up to Waverly station.
I've had a lot of people tell me to start using MDN, but--don't hate me--the thing is I generally find W3S easier to use and more useful (at least for the basic stuff I usually need to Google for). I hate to admit it but I continue to use W3S even if it is at least partially because it's usually the top result, even when I make the effort to scroll down to the MDN entry I'm usually rewarded with something tangentially related and/or missing the actually thing I needed to know.
That said I think there is definitely a lot of room for improvement, I just haven't found MDN fills that gap yet so I wish the author of this luck but I hope they realize MDNs folly. Ad free isn't necessarily better. W3S is, despite it's flaws, fairly simple, easy to use and understand (mostly) and the relevant results almost always show up at the top of a Google search.
I don't think one should be hated for opting for the quickest resource. W3S arguably does have a better design. Some people look for examples in finding out how to code something and as you mentioned, W3S is almost always usually at the very top of the page with browser compatibility at the bottom.
It would be interesting to test out a feature where the user could select simple vs. complex document layout.
This is part of why I think I've developed a reflex, after searching Google, to skip over the first few results after the sponsored links and start looking near the middle of the page. W3Schools, Wikipedia, and a few others. And it's a great example of the central failure of the pagerank idea: if the strongest signal is popularity measured through linkage, the highest quality results will rarely be at or even near the top. Right now Stackoverflow is good and deservedly ranks highly, but I fully expect it to be supplanted in Google's search rankings by an inferior copycat within a few years: one that just happens to generate more revenue for Google by carrying more advertising.
> W3Schools, Wikipedia, and a few others. And it's a great example of the central failure of the pagerank idea: if the strongest signal is popularity measured through linkage, the highest quality results will rarely be at or even near the top.
Actually, I find that the top ranked things usually have what I was searching for -- often that's W3Schools or Wikipedia. I'm not a big fan of W3Schools site design, and I wish they had output examples, but, really, the site design isn't bad enough that it makes it unusable as a reference, and usually when I get it at the top of a search results page, I'm usually looking for a quick reference.
I know that is what they (you?) claim, and I don't have any real evidence to the contrary. The contrary hypothesis, however, explains some otherwise puzzling behavior in Google's search results. So I would like to know if you can offer any evidence for your claim (statements by Google employees are not evidence).
If you are paying attention to the subject of the article and most of the comments on this page, you should have no trouble generating one yourself. Just search for anything involving basic html or css, like, say "html title tag", and see that the first result, and often the first three results, are to an inferior site (w3schools) sporting Adsense ads. Then there are some more like that. The high quality results (w3.org, mdn) start halfway down. None of these have any advertising, and so they generate no revenue for Google.
Half of the people in this thread say that they like w3schools, or even if they don't like it they use it all the time. That is sufficient to explain why it ranks well. Now imagine that the average person issuing these queries is less sophisticated than the people commenting in this thread ...
Among people here who have commented on the comparative quality of these sites, it is almost unanimous that w3schools is far worse than the others I mentioned, to the extent of it actually being dangerous to use. Your reply amounts to a claim that w3schools outranks the higher quality sites in Google's results because it has higher pagerank. Frankly, this is the expected reply. If that is the whole story, it confirms my earlier claim that there is a fundamental flaw in the Pagerank algorithm that conflates popularity with quality. The hypothesis that potential revenue is in fact a signal influencing search rankings is still very much a contender. Not to belabor this, but the protestations of Google employees should not be considered as evidence one way or the other in evaluating whether this hypothesis explains the data.
There's a balance between popularity and quality that we try to be very careful with. Ranking isn't entirely one or the other. It doesn't help to give people a better page if they aren't going to click on it anyways.
You seem to be saying that there is something besides "quality" that influences the ranking of search results, which I find surprising. I thought the idea was to return the best quality results, but there may simply be an issue of semantics at play. When you say "popularity" here I suspect you have in mind something different from what I had in mind when I used the word in previous comments. Can you explain what kind of popularity you mean here (how is it measured).
"It doesn't help to give people a better page if they aren't going to click on it"
I think you've lost me. Don't people tend to click on the top result? And isn't the idea to put the highest quality results at and near the top so people go there? How could it possibly be helpful to the user to not offer them the best quality results?
I looked at the link you provided, but I'm as confused as the author. Isn't there a limit of three Adsense units on a page anyway? As the author points out, Google suggests to publishers that they use the maximum number of ads, and specifically (see the heat map) that they put them above the fold. Then it appears they've decided to penalize publishers who follow this advice. The only way for this to be consistent is if they're only penalizing pages displaying competing advertising products.
The author also shows a screengrab of Google's own results page, where we can see that everything visible on the monitor is sponsored content, with no organic search results at all.
Frankly, I'm perplexed about what point you were trying to make by suggesting this link, but it was interesting.
Almost anyone would agree that the mayoclinic result is higher quality. It's written by professional physicians at a world renowned institution. However, getting the answer to your question requires reading a lot of text. You have to be comfortable with words like "Nonsteroidal anti-inflammatory drugs," which a lot of people aren't. Half of people aren't literate enough to read their prescription drug labels: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1831578/
The answer on yahoo answers is provided by "auntcookie84." I have no idea who she is, whether she's qualified to provide this information, or whether the information is correct. However, I have no trouble whatsoever reading what she wrote, regardless of how literate I am.
That's the balance we have to strike. You could imagine that the most accurate and up to date information would be in the midst of a recent academic paper, but ranking that at 1 wouldn't actually help many people. This is likely what's going on between w3schools and MDN. MDN might be higher quality, better information, but that doesn't necessarily mean it's more useful to everyone.
Nobody denies that W3Schools is a terrible place to learn how to program. Their examples are outdated, insecure, and don't really explain why things work in a certain way. Their tutorials are optimized for copy & paste, and God knows how many security bugs in the wild are due to somebody copying W3Schools.
But W3Schools also carries some of the most easy-to-use indexes of pretty much everything related to web development. Can't remember the exact name of that CSS property and its possible values? Can't remember the order of arguments of a JS string/array function? Go to W3Schools and browse their lists, get the answer in 5 seconds.
Of course, there's Mozilla Developer Networks, too. But MDN's CSS Reference  is just an alphabetical list of CSS properties with no explanation. W3Schools' CSS Reference  is neatly categorized, and each entry has a short explanation next to it. Very useful when you're trying to remember, for example, "word-wrap: break-word", which is not the same as "word-break: break-all".
W3Schools is good ol' Web 1.0 stuff. Just some lists and a bunch of links. No need to ask somebody and wait for an answer. No need to watch any annoying videos. No need to follow an animated tutorial designed by someone who thinks they know how a person should learn programming. W3Schools is like an old dictionary. A dictionary isn't particularly helpful when you're trying to learn a foreign language properly, but it's quite handy when you're face to face with a foreigner and you need to think up a word for "toilet paper" quickly.
I use W3Schools when I need to remember something quickly, and MDN when I need to learn more deeply about it. When I write CSS, I always keep W3Schools open in one tab and MDN in another tab. I wish somebody would link W3Schools' lists to MDN's details. That would make the best of both worlds.
Above the fold, one is a very quick and easy way for users to copy code and modify it for their purposes. It's also stupidly simple language. To make a box shadow, do this. The other has undisputedly better technical data but forces you have to scroll below to just see a sample code. And most new coders wouldn't understand how to read their syntax section. On top of that, w3schools lays out values in tables whereas MDN has paragraphs of copy.
By no means am I arguing that MDN doesn't have more detail, but I can definitely see why W3Schools is still used by those who know better.
Really? I just did a quick comparison, and w3schools was much better than MDN. Try looking for info on the 'required' attribute of the input tag. w3schools has all the attributes in a compact table making it easy to find the attribute you're looking for at a glance, whereas MDN has a big long list. In w3schools you can then click on the 'required' attribute to get more info about it and see which browsers it works on. In MDN there is no info on browser compatibility in the 'required' description - you have to scroll down to ANOTHER table (there is not even a link to it) to see what browser it is compatible with! Come on, even back in web 1.0 we had links to stuff!
The reason that w3schools is at the top of the search results is that more people find it useful that MDN. Perhaps it isn't as cool with the hipsters as MDN, but for people trying to get things done w3schools is more useful in my opinion.
The hipster comment was related to the fact that the article is complaining about ASP, "tight layout" and "icky". I don't use ASP myself, but it really doesn't have any bearing on the content. I don't mind the ads myself (they're fairly unobtrusive), and the layout is acceptable.
I like nice designs myself (I'm currently building a new site and I'm getting inspiration from roon.io and other beautifully designed sites), but I'm also quite content using sites that have less-than-ideal designs if they do the job that I'm looking for.
I've never heard of mdn before this thread somehow. But w3schools is perfect for looking up the things I tend to look up (mostly CSS or HTML attributes) and it's really easy to parse. Plus it's always the first google result for me.
I've been developing websites since 1996 myself, and MDN hasn't really been on my radar for an html reference. I did use it extensively back in 2009 when I was doing work with <canvas>, and back then MDN didn't really have a general html reference. I think it's only in the last couple of years that MDN has really developed their reference documentation.
So, if you've always used MDN you really must be a young whippersnapper :)
I think MDN has an advantage in that they've only really been around since HTML5 has become standardized, whereas w3schools has had to deal with the changing standards of html4/xhtml/html5.
"I think MDN has an advantage in that they've only really been around since HTML5 has become standardized"
=> HTML5 is not a standard. Nowhere near so. Whatever you mean by that, MDN is as old as Firefox anyway.
"...whereas w3schools has had to deal with the changing standards of html4/xhtml/html5."
=> These differences are 99.99% additions, so don't impact documentation that much. In any case, documentations is more relevant is it documents implementations, not standards (which come after implementation most of the time, not the other way around until recently)
You can do that with Firefox's bookmarks (last I looked) and Chrome's "search engines" (and there's probably something in Opera or Safari) using any URL you want, %s in the URL will be replaced with your search.
Indeed. I'm always ashamed of using it, but I have a stupid tendency to forget css selectors from time to time (I don't do web design or tweaking that often) and the same goes for how string.match or regexp.test or the likes in JS.
For me it's not only web development, is most of the CS stuff I need when coding.
Probably it is due to my training as a mathematician, but I usually just keep an index of what I know and what it's used for, but I don't remember the specific details. I.e. Implicit Function Theorem can be used for proofs involving X, Y and Z under hypotheses H, unless you are fancy and go for H2 and more work. I just know what I need and where I can look it up – Nirenberg for fancy, any basic Differential Analysis text for the others, that Russian book on my shelf for a neat simple proof, Moser's for a simple proof and application, the list goes on with 6 or more 7 items. Well, this is actually stretching it a little thin, I could prove the basic differential versions without much trouble, and some functional versions given enough hypotheses.
For a long time I always forgot the order of the 4 arguments of the margin/padding properties (it's top right bottom left). I always used w3schools to look it up. I use w3schools just to lookup attribute and CSS property names/values etc. It has a nice tree at the left side and concise tables of possible values. If anyone would want to prevent me from using w3schools he/she has to replicate this. It could use some updating and better information about browser support (maybe from a shared source with caniuse.com). I didn't even know that there are tutorials and such on w3schools.
Agreed. Talk is big, but this person hasn't done anything. Heck, they could have thrown up a wiki and written a little one sentence starter page about each CSS tag in the time it took them to write the post, then ask for contributions.
Perhaps to their credit, they have a two-part tutorial on email sending, one titled "PHP Email" and the other "PHP Secure Email". Also, their form handling tutorial is followed by another tutorial on "validation" (by which they seem to mean a combination of HTML escaping and outdated defenses against SQL injection). But most people are just going to copy and paste the first tutorial without looking at the second.
I can't speak for anyone else as I don't and have NEVER used W3Schools for anything related to PHP or MySQL but used PHP's manual and MySQL docs for anything, but when someone needs help and asks me, I usually use my experience and write up a quick example and do not go through the whole sanitizing of inputs and that, as they're just an example and sometimes are written right on the spot but I do let them know it's only an example and not to be copy and pasted into their documents and that they should properly secure their stuff with input sanitization.
I really love MDN, and in some ways I feel guilty for doing this. I've been a MDN contributor.
But the MDN information architecture/navigation is terrible, the crowd sourced content is inconsistent, the pages load very slowly. When you put it all together, it's difficult to focus on winning the search engine game with those flaws.
There was a "Better JS docs" movement a few years ago that moved the needle a little, but in the end it was just the wrong content to do the job.
W3Schools is nothing more than a simple to use (google) and navigate (google) collection of pages with tons of reminders to jog your memory about what exactly it was that a specific css selector or html tag did again.
If you use it to learn PHP programming you're doing it wrong. If you use it to learn MySQL, you're doing it wrong. If you use it to learn how browsers work, you're doing it wrong. As a Cliff's notes for HTML and CSS, however, it is more than adequate.
I have a confession to make and a risky one, well to the extend of every karma I got. I own http://www.w3fool.com/, the one with missing 's' and you don't want know where it redirects to. I bought it when I saw it at an expired auction. I couldn't stand at thought of a website existing just to shame another.
On one hand, I agree with most of you that W3School isn't accurate or up-to-date or is riled with example that have security holes but on the other I am really thankful for them for getting me into web development. Back then there weren't much options, W3School was the easiest and the best we had. I remember sitting late nights learning to build simple web apps by referencing W3School every few minutes. Getting someone interested into something is very hard, to get a kid interested in sports, you don't start with a ruleset, instead you give him a start and correct him as he makes mistakes. That start was what W3School was for me.
Although I have been defending it, I think they should expire/update w3school. Times have changed and there are better resources to take its place.
I was expecting to see some examples about why there are so much hate in W3schools but I didn't find any on this rant.
I remember to have seen some bad examples showing bad habits but to be honest, I have seen more people ranting about W3S than bad examples on the website, maybe it would be more useful to show why W3S sucks that much.
Members of W3C have asked W3schools to make it clear that there's no link between W3C and W3Schools, but W3Schools declined.
W3Schools offer "certification" - (http://www.w3schools.com/cert/default.asp) - of dubious value. This is troublesome because people know of W3schools, and sometimes think there's a link between W3schools and W3C.