Is the general quality of Web Development diabolically low?

Get Started
1

Is the general quality of Web Development diabolically low?

While looking for good companies in the UK to apply to work for, I found a list of the ‘Top 100 Web Development Companies in the United Kingdom’ here:

Obviously I want to work for a company that does genuinely good work, so I thought I would run the first and most basic checks on these companies; Whether their home page passes W3C validation for CSS and HTML.

Here are the results:

[table=“width: 500, class: grid”]
[tr]
[th]Company Website[/th]
[th]W3C HTML Validation[/th]
[th]W3C CSS Validation[/th]
[/tr]
[tr]
[td]http://www.holbi.co.uk/[/td]
[td]Fail[/td]
[td]Fail[/td]
[/tr]
[tr]
[td]http://www.fiveriversonlinemarketing.co.uk/[/td]
[td]Fail[/td]
[td]Pass[/td]
[/tr]
[tr]
[td]http://www.nvisage.co.uk/[/td]
[td]Fail[/td]
[td]Fail[/td]
[/tr]
[tr]
[td]http://www.boxuk.com/[/td]
[td]Fail[/td]
[td]Fail[/td]
[/tr]
[tr]
[td]http://www.clickconsult.com/[/td]
[td]Fail[/td]
[td]Fail[/td]
[/tr]
[tr]
[td]http://www.virvo.com/[/td]
[td]Pass[/td]
[td]Pass[/td]
[/tr]
[tr]
[td]http://www.omdesignlondon.co.uk/[/td]
[td]Pass[/td]
[td]Fail[/td]
[/tr]
[tr]
[td]http://www.newbrandvision.com/[/td]
[td]Fail[/td]
[td]Fail[/td]
[/tr]
[tr]
[td]http://www.redweb.com/[/td]
[td]Fail[/td]
[td]Fail[/td]
[/tr]
[tr]
[td]http://www.dotsquares.com/[/td]
[td]Fail[/td]
[td]Fail[/td]
[/tr]
[/table]

Unless I am missing something, this is a really bad sign, that 70% of the these companies fail both CSS and HTML validation, and 90% fail both HTML and/or CSS Validation. These companies are making a lot of money, customers trust them as experts in their fields, but they haven’t even bothered making their own homepage pass the basic W3C Validation Tests.

What does the community think? Is the general quality of Web Development diabolically low?

2

[font=verdana]Much as I am a standardista who wants to see pages validate properly, you need to be careful about saying that pages “don’t validate” … there can be different reasons why. For example, a lot of people have social media widgets embedded in their pages, and these often use technically invalid code such such as unknown attributes, which are harmless to browsers. Or people are using newer CSS properties with vendor prefixes, which will again cause the validator to fail even though they won’t give any problems to users. I wouldn’t count those things as a “validation fail”, even though the W3 Validator will throw up errors.

There is a lot of bad practice out there, no doubt about that, but I do think the situation has been improving over the last few years, as more webbists have become switched onto accurate and semantic code.[/font]

3

That’s interesting. But if the page doesn’t conform exactly to W3C guidelines, then can you really expect the page to display and function correctly on all W3C compliant browsers? I mean can it effect people using less conventional browsers, such as assistive technologies? The whole point of everyone working to a predefined standard is to prevent compatibility issues.

4

The web has moved forward whereas the W3C validator hasn’t quite caught up yet. You can have perfectly valid CSS and HTML sites that are completely inaccessible, unusable and a genuine disaster, yet they’ll pass the technical validators without problems which is why I’d never use validator tools as means to measure the quality of a website’s design.

The only way to find out whether a site is good is by looking yourself, evaluating the content, scrutinizing the code, doing some tests on performance, seeing how a site works on various devices, modern browsers, legacy browsers, with scripting turned off, images turned off, etc.

If I had to evaluate the quality of a site, here’s what I consider important checkpoints (in no particular order):

a) Good color contrast between foreground and background (text, images, media)
b) Optimized images (as low kb as possible) + fallback (good ALT text where apt)
c) Legible text
d) Good content (grammar, spelling, language, form)
e) Good navigation
f) Logical content structure
g) Good, well-organized and slim code
h) Easy to use on all devices
i) Works in all modern browsers, functional and decent in legacy browsers
j) The design works with Javascript turned off (on sites where applicable)
k) The design works with images turned off

What I mean to say is, that by using the validator, you’re addressing just one single point of a perhaps dozen that are needed to make a proper judgement, imo, as, like Stevie D said, markup errors in an HTML document can occur extremely quickly, be it that a content editor made a typo, a CMS with hiccups in the markup output, a backend script/add-on that has markup hardcoded in its functions (happens often, unfortunately), and other factors that may lead to a site’s markup not validating. And for CSS3 as well as vendor-prefixes, the current standard validator is absolute rubbish as it only recognizes CSS2.1, yet the browsers have since long moved on and quite a few of the new CSS3 features are indeed cross-browser compatible, even in some legacy browser versions.

5

Couldn’t have put it any better than Maleika (kohoutek) just did, but all of the above points to the fact that validation is just a means of flagging possible problems rather than a blanket test of a site’s worth. A validator will warn you about vendor prefixes, as mentioned, when they are perfectly fine to use. And even though your doctype may not like certain deprecated elements, they still tend to be fully supported by browsers anyway … etc.

6

Thanks, that’s interesting. So in essence, the validators need to improve, and should only be used as guidelines. I know, of course that it isn’t the only thing to look at, but I considered it the per-requisite. This is because I thought if the site didn’t produce markup that passes W3C guidelines then there would be a whole load of stealthy but critical problems that would pop out of the wood work. So there wouldn’t be much point in looking further if the basics haven’t been achieved, it would be like a really nice car without any wheels.

So if you should use the validators as a guide, the question I have to ask, is when do you ignore the validator? Especially, how can you tell when a validation error effect SEO?

7

A better comparison would be a car that may work just fine even if some of its parts are not those recommended by the manufacturer.

when do you ignore the validator?

When your site works fine in all browsers. :slight_smile:

how can you tell when a validation error effect SEO?

I’m not sure it would all that often. The best thing to do is make yourself an SEO checklist, and compare it against any code issues you have, and see if there’s any overlap.

8

By saying test your site until it works in all browsers, I assume you mean just the most popular ones. In which case, should we just ignore the minority browsers and hope for the best? Screen readers for the blind can be especially difficult to test (because you have to listen to everything before you can determine the web page works properly) unless you are used to them.

9

Yes, it’s not feasible to test in all browsers, as there are too many of them. But test in as many as is feasible. Testing screen reader performance is a tricky one, because it’s hard to replicate the user experience unless you are a screen reader user (and everyone presumably uses them differently, anyhow). The best (and probably only valid way) to test screen reader experience is to ask a group of screen reader users to do the testing for you. Aaaaaand just hope you’re a big company with mega$$$s.

10

It’d be hard to test-drive a car without wheels though. :slight_smile:

The W3C validator site itself has a good roundup: http://validator.w3.org/docs/help.html#validation_basics

If a site validation fails, it doesn’t mean that the basics haven’t been achieved. Let’s say you have created a site, have added no CSS3 or vendor-prefixes but standard CSS2.1 and standard HTML, all of which the validator accepts and passes as valid. You have added a CMS, the markup is output by that CMS. All good so far. You hand over the site to the client. The client is or has one (or multiple) human editor(s)/writer(s). These editors must use some kind of text editor or markup aid like Markdown or Textile or something else. It may very likely happen, that they screw up something, maybe forget to open a <p> or encapsulate a <p> into another <p> or something else.

With just that tiny mistake, the whole page becomes invalid. If this mistake does not render badly on the site, meaning it doesn’t break the content flow or page layout, then that’s nothing I’d consider harmful or bad. More importantly, it doesn’t lower the site’s quality for me.

As for SEO impact, the markup needs to be highly broken for search engines not to be able to read the output. I have no proof or evidence for that, though. I do know, however, that all the big and smaller news sites and magazines I read are full of markup errors and yet their content ranks very highly in search engines regardless.

11

A car can only be good if it uses “traditional”, “approved” car parts? Where do the innovations fit in?

And, like a car, a web page is only as good as the user satisfaction factor it brings. As others said, they don’t build cars to satisfy the car industry regulators. There aren’t enough of them to make it profitable.

12

It seems like there is a gap in the market for a validator that can deal with the times.

13

You can, but it’s a bit more work: http://www.impressivewebs.com/css3-validation/

14

Here’s some sites that could be considered successful, yet don’t validate.

youtube.com

yahoo.com

amazon.com

Google and amazon no doubt care about their users experience, and validation hasn’t won out amongst dozens of other measurable attributes as a priority.

Content, utility, usability are all massively more important, as are all the other things in kohoutek’s list. Getting hung up on validation as a measure of quality isn’t a worthwhile endeavor.

15

I don’t see web design or code development in the same way. To me it is the same as art, with various mediums and tools at the artists displosal. There are successful artists who’s work you do not like, and there are poor artists who’s work you may love.

There is only your right way for you

A strange question this… Is the general quality of art diabolically low?

16

Presentation should be separate to content.

17

True from the perspective of the coder. Has absolutely no relevance at all to the end users experience. It’s easy for coders to get caught up in only the precision of their craft, and think that’s all that matters.

Do you create web sites for people to use and enjoy, or as an exercise in your own coding perfection? Which is more important?

What kind of paper was the mona lisa painted on, what kind of paint? You can tweak pigments and oil and papers but never paint a worthwhile painting.
There is of course worth in knowing how to mix pigments and oils, but an expert in such that can’t appreciate the final form of beauty has absolutely no qualification in being an art critic.

18

Is the general quality of Web Development diabolically low?

Yes, but you’re missing the overall point. Faceook to my eyes is an awesome website, it not validating does not render the website as having low quality web design. Many sites don’t validate because as @Stevie_D; mentioned those pages might have addition widgets or be implemented on a CMS which prevents clean validation, not that this matters extensively to the overall performable of the site.

Another issue to raise is that the template, would validate, however once implemented and made functioning, the LIVE website won’t. Having a valid template would allow your web page to run successfully in all browsers.

Just to illustrate, putting a simply Facebook ‘like’ button or a YouTube video would screw your validation up. For those companies it’s more about having the function works in ALL browsers quickly and efficiently, rather than validate with w3c standards of which they’d struggle to know which one of those standards you chose to adhere too. As of such their code should work in all doc-types irrelevant of whether the code validates or not, as they focus their attention on function.

I don’t see web design or code development in the same way. To me it is the same as art, with various mediums and tools at the artists displosal. There are successful artists who’s work you do not like, and there are poor artists who’s work you may love.

There is only your right way for you

A strange question this… Is the general quality of art diabolically low?

Yes!

In my experience the art of web design is very low and highly unregulated. I must say those the examples @ro0bear ; showed above do not demonstrate that. The industry is flooded with too many people with too little experience not knowing what they are doing, furthermore selling those services to businesses. I can’t say much, as I used to be one of those people myself. :stuck_out_tongue: The ones who stick at it make a career, the others who struggle to make a living change their career path.

It’s also important to realize the purpose of the website, and what it’s goals are. There are so many variables in a site that we’d struggles not to get carried away into it’s those minor detail.

A client from my experience would not have the validation of the code as a priority. Instead they will look at the sales generated and also as importantly, how the users (new and current), interact with their website as this too could affect sales. I hate to say this, but the awesome design we might have coded and made work comes secondary to the generated revenue from the website.

19

@Saga I see what you are saying. I do agree that the customer usually (almost always) values sales before correct markup, but if your site does have correct markup, it will be much more future proof because the code will be based on design standards rather than being hacked to look right in the currently popular browsers. Also, I don’t think correct markup needs to be at the cost of sales (apart from perhaps some widgets).

@Saga, @EastCoast, @kohoutek, @ralph.m, @Stevie D
I have thought of some possible solutions to compromise between correct validation and keeping with the times. I am assuming you agree that correct validation in an ideal world is a good thing.

Option 1
Code your site so that when testing, a variable it put into a cookie which the server side language reads. As a result the program cuts out the widgets etc from the markup. As a result, your page should validate cleanly, this ensures that everything is correct when the widgets are not present.You can then use a firefox validation plugin to view your pages which will flag up any markup mistakes.

Option 2
Don’t hard code widgets into the markup, instead add them in front end some way using JavaScript, perhaps just adding them directly, perhaps having a hover over button to bring up a list of widgets you can access etc?

Option 3
Have a link to a page with the widgets on, rather than having them on every page.

@kohoutek Canvas and paint is a medium that communicates directly through the physical world to humans which interpret it. With a computer, the art has to be interpreted technically before it is displayed on the screen. That is why it is important to communicate to the computers correctly if you want the art to be communicated correctly to the end users. It’s like a game of Chinese whispers, the programmer tells the computer, the computer tells the end user and if the computer misunderstands, the end user will get something completely different to what the programmer intended. You have to take into account that the information is not only communicated visually, there are people who interpret the content only audibly, and those who interpret it only by touch, and those who try to access it via less conventional means such as an obscure phone browser, or a search engine scanning it for indexing. If you want to show good art, you have to ensure the canvas is reliable, otherwise even the best artist can fail.

<offtopic>
p.s.
how do you do a mention link?
</offtopic>

20

As mentioned already, validators are a poor metric of quality. At least the W3C validators. Semantics are far more important, IMO. Using tables for data, etc.

In the world of PHP the quality of code is the worse I’ve seen:

  1. There is no standard framework (Zend is defactor sure but not that common)
  2. Low learning curve means limited experience developers can begin work right away
  3. Patterns and principles seem to observed and attempted but often patterns are difficult to comprehend in the context of a request/response environment.

Quality of code unfortunately, has little bearing on the opinion of users of software. WordPress, last I checked would fail most quality assurance checks, but it’s a huge success.

Just my two cents :slight_smile: