November 09, 2005
To my recent post on semantics, a couple of comments touched on the wider issue (and one I have never really understood) that somehow by standardizing, you stifle creativity. It's an argument I've heard for many years in criticism of web standards, and one that I fail to see any logic in. But I've never really seen it discussed in detail, and as it simply won't go away, I want to address it here.
In a comment on Semantics in the Wild, hostyle wrote
If everyone uses the same tags, the same attributes and the same attribute values, guess what you'll end up with? Either:
a) An entire internets worth of the CSS Zen Garden?
b) A boring bland internet where design can on longer get around the mediums limitations?
Which do you think is more likely? Are there not enough bland default-styled blogs out there already?
I think this continues to be one of the great furphys (a kind of red herring, apparently a term unused in North America?) regarding standards, formalization, and the web.
The argument is that standards are bad, because somehow standardization will stifle innovation and creativity. I frankly can't see how that logic works.
All communication media (other than the web) are rigorously standardized. Radio, Television, film, sound recording, all have strict technical standards. This brings enormous economic benefits. While much of the output of these media is, I would be the first to agree, repetitive, uninspired, uncreative, it can hardly be blamed on the fact that if you buy a CD it will work in your CD player. Could someone explain to to me the logic that if somehow you had to buy specific CDs for each specific type of CD player you owned, then there would be more creativity in music?
But that is argument by analogy, which is rhetoric, and so, not very useful.
Let's take the proposal that we somehow standardize the class and id values used for common information patterns which developers are already using and have done for some time. hostyle and others argue that somehow this will lead to "boring bland internet where design can no longer get around the mediums limitations".
But, what I propose is built on top of a considerably more impoverished semantic framework, namely HTML - HTML has a very very limited set of elements with which to work. Surely the argument should follow that we need to get rid of HTML, and simply let people make up their own web languages to encourage creativity?
Constraints are a fact of any medium of expression - language, music, cinema, whatever. In fact, you might argue they define a medium. Is the music of J. S. Bach, constrained as it was by both conventional and self imposed rules, any the less creeative for that? The constraints of the fugue, for instance, define much of the music of Bach, among the most sublime, revolutionary, and indeed popular in western music.
Or am I still missing some vital piece that eludes me?
November 9, 2005 | Permalink
TrackBack URL for this entry:
Listed below are links to weblogs that reference Web Blandards:
You aren't missing anything, I don't think.
The idea that standards leads to stagnation is rubbish. Standards allow unrelated things to be used with each other - like the examples you used. We just take standards for granted, cos they *just make stuff work*.
Standards only exist, though, because they make things better than if they *didn't* exist. If there were lots of versions of CDs, or of electrical appliance voltages, or of nut and bolt sizes, then we'd all be a bit screwed. But I'd guess that those standards emerged from a period of variability, of 'creativity' even, out of which rose a direction on the best way to do things. Then those ways were adopted more widely until they were accepted as the standards.
The key to the standards is that any drawback of using them (prevention of finding new and amazing sizes of nut and bolt to use) are *outweighed* by the benefits (the ability to use any given tool to tighten them).
The benefits of using valid XHTML (etc.) are that any compliant browser can render the page. The days of cross-browser problems are *gradually* coming to an end (when and if they will is another discussion). Getting creative and inventing your own web language could be great, but the benefits of interoperability brought about by standards outweigh that.
The big question for this debate is: what benefit comes from standards in the naming of classes and ids? Apart from the glowing feeling that standards are somehow inherently good, independently of any other consideration?
The answer is maybe along the lines of the idea behind 'The Semantic Web', which I don't know enough about, but which Tim Berners-Lee is always discussing. If we all used the same attributes to name the same kind of information, then we could get at that information in new ways.
In my job, I put lots of information into webpages - say, things like dates and times and venues for lectures. If I used ids in common with all other dates, times and venues for similar things, then someone else could reliably write a system which crawled the web finding all the lectures on a given date in my local area. Alternatively, a browser could have built-in integration with my calendar, and it could grab the date (and other) information straight from the page of the lecture. Very goddamned useful.
It's certainly more complex than this, and this is certainly not all there is to it (and I could just have lots of this all wrong...) but the *idea* of standardisation of this kind seems to me to have possibilities that are near-enough endless. Rather more than just CSS Zen, which is a great example of the same thing, but applied only to presentation (a trick which is definitely useful in itself - imagine being able to reliably switch into 'show me only the content' mode in your browser...), a real semantic web is *better* than what we already have.
It won't work, as John says, by just coming up with a list and suggesting that everyone do it. It's all about the cowpaths, or desire-lines, or whatever you want to call them. These things will emerge if we start looking for them.
In the meantime, you can, of course, use standards in a boring way... and we've all seen a lot of people who do that every day.
Posted by: rich | Nov 9, 2005 10:00:33 AM
rich, well put.
I agree with John's original comments on standardisation NOT leading to stagnation and you sum it up very nicely. Just because we may all use similar class or id names one day doesn't mean that we have to implement them the same way. From a design point of view, take an id of mast for your banner for example. You can implement that in a myriad of colours and dimensions with any background image you choose. No stagnation there as everyone will implement differently.
I don't think stagnation is a major issue, so there may be a little stagnation from a design point of view. The benefit as you mention of a commonality that leads to added functionality being available or developed far outweighs any potential stagnation.
Posted by: damo | Nov 9, 2005 11:38:37 AM
Great comment Rich. An aside:
I can't help noticing the lack of standardisation of the spelling of standardization! John, I take it that even though you are Australian you choose to spell this word the American way, and then two commenters choose the contrary. Ah well, small things, small mind perhaps!
I don't see a problem with the lack of standardisation of ID's and CLASSes as to force people to use a similar framework would cause some square peg / round hole content shaping that anyone who has used a CMS would recognise as quite limiting.
Posted by: Adam Bramwell | Nov 9, 2005 1:13:42 PM
Rich and Damo, good points, thanks for that.
Adam, the point is not to force people to use inappropriate solutions to their problems, rather, to help solve similar problems in similar ways. I amworking on a much more detailsed post to outline this in more detail, particularly the benefits you might get from this approach,
Posted by: john Allsopp | Nov 9, 2005 1:23:23 PM
As someone coming originally from art school, then film/TV and animation... and finally (visual/interface) web design I would have to agree at a pactical level. I don't think a notion as described by John can be argued against in terms of real-world practice. Constraints and limitations are too often mistaken as bad things, when in fact in the creative domains its often these very constraints that lead to innovation through invention.
It's now evident online that web standards doesn't hinder visual creativity, so the arguent is clearly about technical constraints. Did the crappy VHS standard stop us doing useful and clever things? In fact, how many new standards have we had since those days?... and still it will keep evolving. Same on the web. Don't complain, create!
Sure HTML and its offshoots since the early nineties has spawned hassles, but as John suggests thats no excuse for lack of creativity,.. perhaps those who complain about the tools that lack in turn 'lack' creativity themselves? (Spend some time outside your hotel in the back streets of a poor country and you'll soon see what being creative with serious constraints mean). The current web world out there is an undeniable reality (and don't forget the historical context for it), its the 'paintbox' we've got regardless of the reasons why its a problem.
Meanwhile, all the valuable creative reactions against currrent 'standards' will yield yet another improved 'standard', and still there will also be those who grumble and complain - it will never stop. Perhaps that is the nature of the innovation cycle.
Posted by: Andrew Francois | Nov 9, 2005 6:58:48 PM
I can see some sort of reasoning behind this "standards limit creativity" statement. For example, lets take four ID-values commonly used on pages: header, footer, nav, and content. Maybe those names are a bit too unsemantic, but lets say those names are now standard names, you should assign to the specific regions of your site (if you of course have those regions).
This seems rather limiting, because where else would you place header, than somewhere at the top of the page. And where would you put the footer, probably at the bottom. You can play around a bit more with nav and content, but you wouldn't put nav at the center of the page and push the content into thin right column.
The thing is, it's not the standardised markup, which prevents you from doing that - it's the common sense!
And when we look at the web pages around the world (for example like François Briatte did in he's survey "These web sites are identical—or are they?" http://phnk.com/design/survey/ ), we see, that most of the web sites really have quite similar layout. But this isn't neccesserily a bad thing. Actually we don't want to make pages too different, because then our wisitors (who have spent the most of their time not on our website) would get lost and possibly leave our carefully designed site.
And when we look at other forms of media: newspapers and books have headers and footers and a title page, movies start with title and end with a list of actors... and this is what people really have come to expect.
Posted by: Rene Saarsoo | Nov 9, 2005 8:57:12 PM
An aside - I have looked up the standards for the word 'standardize' :-)
First, apparently, where there is the choice between '-ise' and '-ize', '-ize' is the preferred ending. This is the American way to do it, but it is "endorsed by the Oxford and Cambridge University Presses' (according to the Penguin book 'Usage and Abusage', a great book which runs through all these kinds of things.)
But according to the BBC, "The rule used by a number of British dictionaries is that words with a Greek root take the -ize suffix and those with a Latin root take -ise. This explains the large number of -izes in a British dictionary".
However, the BBC itself seems to use 'standardise' - so I'm a bit confused. I am from the UK... do I follow the BBC, or the dictionary, or what?
I'm going to get creative and start spelling it 'standard-eyes'.
Sorry, back to the subject.
Posted by: rich | Nov 9, 2005 9:23:41 PM
Whilst we are kind of offtopic (thanks rich). In a document I was writing up at work today I was spelling 'colour' as 'color' and didn't even notice it until the spell checker picked it up. Stupid CSS!
Anyway, I am from Australia like John and I have always spelt standardise with an 's' so who knows :-)
Posted by: damo | Nov 9, 2005 10:36:47 PM
To broaden the perspective on this discussion theres an interesting PopTech podcast of Barry Schwartz on the topic "Less is More" and the idea of too much choice being a bad thing as much as too little choice is a problem.
"Standardizing" is a process of limiting choice, not in order to limit creativity but in order to maximise the benefits of and access to the thing being standardised to as many people as possible. This is surely the aim of web standards.
This feeds into this idea that in evolutionary terms of the 'standards' (to take an ecological view of the web) there's a spectrum where total constraint/limitation or infinite possibility work at opposing ends to determine a 'sweet spot' or balance where returns/benefits are optimised. Standards in this sense evolve from a process of 'natural selection' by the developer AND user community in a very 'democratic' way... but like most evolutionary processes it is slow. Luckily evolution on the web is measured by months rather than millions of years.
Posted by: Andrew Francois | Nov 10, 2005 11:46:29 AM
Great comments, thanks for them all. As you can see, I have been busy updating my site - hopefully it dosn't look too blando.
Posted by: John Allsopp | Nov 10, 2005 3:59:18 PM
One of the first standards was the Roman 'Passa' or pace, roughly equivalent to our 'yard'...
...and to paraphrase...
..."what did the Romans ever create for us?"
Lovely article, well written.
Posted by: jim | Nov 11, 2005 3:20:25 AM
To me the most important creative aspects of a site are determined before the first keystroke of code is struck. Reusing markup and CSS reduces time spent coding and dreaming up id and class names. Thereby creating the opportunity for more time to be spent on aspects of the site for which there are rewards for being unique, such as design and function. Standardise!
Posted by: Ollie | Nov 11, 2005 9:18:00 AM
I'm glad that you're carrying on this discussion and taking all views into account. I never claimed that standards are bad, and my main point in that comment was the last line:
> CSS is for visual style and presentation, not semantics.
You suggest standardising on common class and id names, but what you're doing IMO is going down completely the wrong route. If you wish to gain more semantics from existing standards, I would suggest using microformats and expanding on XHTML tags and/or attributes using custom DTDs.
As for stifling creativity: some designs I have had to convert to websites would not conform to your "standard" header / footer / menu / widget-that-goes-over-there list. Did I do something wrong because I do not meet your standard? Or is there something wrong with your standards basic premise? My website did, however, validate with the w3c HTML and CSS validators.
Basically you're suggesting another standard with a very limited set on top of an existing limited set, and claiming that this will somehow make things better? Perhaps you are ignoring the bigger picture - have you ever asked yourself why HTML is as limited as it is? IMO you're blindly limiting this set of tags further by suggesting people standardise their CSS also.
Let people use their own creativity to do as much as they can with the current limited standard. This is of course just my opinion - you can do and advocate as you please. I'll just re-iterate that "CSS is for visual style and presentation, not semantics."
rich: "If we all used the same attributes to name the same kind of information, then we could get at that information in new ways." - this is what semantic XHTML / XML / HTML / microformats are for - not CSS! I think perhaps people are not exploring existing semantics enough, and are already trying to create new ones.
Posted by: hostyle | Nov 15, 2005 10:44:43 PM
>I'm glad that you're carrying on this discussion and taking all views into
I try to converse rather than pontificate. The degree of success is open to debate
>I never claimed that standards are bad, and my main point in that
>comment was the last line:
>> CSS is for visual style and presentation, not semantics.
But of course class and id are part of HTML, and CSS merely provides a mechanism for styling elements based on these attributes (in future, the attribute selector is probably the better way of styling based on class and id values)
>You suggest standardising on common class and id names, but what you're doing
>IMO is going down completely the wrong route. If you wish to gain more semantics
>from existing standards, I would suggest using microformats and expanding on
>XHTML tags and/or attributes using custom DTDs.
I'll address these one at a time.
I agree microformats have their place here. But µF have a very specific focus - they concentrate on data ("a way of thinking about data") (rather than architectural components of a page ). In about Microformats at microformats.org you'll read "microformats are a set of simple, open data formats built upon existing and widely adopted standards", while microformats are not "infinitely extensible and open-ended".
As mentioned in a couple of previous articles, I've had extensive conversations with Tantek Çelik, Rohit Khare (and others) about microformats, spoke at the developer day at WWW2005 in Chiba on the microformats track, and realised that this higher level approach, while complimentary to µF is not appropriate to µF. Above all its a philosophically issue based on the underlying principles and aims of µF.
As for "expanding on XHTML tags" - how exactly would you propose that? And why do something more than developers are currently doing? If we take Einstein's maxim "make it as simple as possible but not simpler", what do we get by adding to XHTML (however you might do that, from waiting several years for the W3C to come up with a solution for you, to custom applications of XML which you then use with XHTML 1.1 and namespaces - which of course lots of developers will do, right?) . That is, what's so broken with using class and id values that you'll fix with custom DTDs and extensions to XHTML?
And, after all, µF is simply using existing XHTML elements and attributes, and to an extent class values, so from an implementation point of view, it suffers equally from whatever shortcomings you identify (but don't outline) with my approach.
>As for stifling creativity: some designs I have had to convert to websites would
>not conform to your "standard" header / footer / menu /
>widget-that-goes-over-there list. Did I do something wrong because I do not meet
>your standard? Or is there something wrong with your standards basic premise?
Perhaps I have not sufficiently articulated my point.
I am not advocating that all sites conform to a particular structure. What I have observed empirically, as have others, is that many many sites conform to a set of patterns, and yet, at an implementation level, they are usually implemented in totally different ways. Take blogs. There are 10s of millions of them, apparently 70,000 new ones a day. They almost all conform closely to a set of patterns in terms of their architectures,and yet, under the skin there are countless permutations in terms of how they are implemented.
I am about to publish a very detailed post about patterns in web development, so I won;t go on further. Except to observe that I am not proposing a prescriptive "you must" formula, rather, a descriptive "if you do this, here are some ways you can solve it that work" approach. Stay tuned for that article, I do hope it clears things up.
>Basically you're suggesting another standard with a very limited set on top of
>an existing limited set, and claiming that this will somehow make things better?
My suggestion has nothing to do with limitations,but again, I must not have made that clear.
>Perhaps you are ignoring the bigger picture - have you ever asked yourself why
>HTML is as limited as it is? IMO you're blindly limiting this set of tags
>further by suggesting people standardise their CSS also.
See the above, I think sadly I have not sufficiently clearly outlined my ideas, hopefully its a bit clearer now.
>rich: "If we all used the same attributes to name the same kind of information,
>then we could get at that information in new ways." - this is what semantic
>XHTML / XML / HTML / microformats are for - not CSS! I think perhaps people are
>not exploring existing semantics enough, and are already trying to create new
I think there is some truth in that observation, but I also think there is a level of semantic complexity that current XHTML compounds and so forth simply can't provide. I'll address this in more detail shortly in the above mentioned article
Posted by: John Allsopp | Nov 16, 2005 7:34:19 AM
Check out the latest AIGA podcast with Mr Zeldman himself: http://www.aiga.org/content.cfm?ContentAlias=jzeldman02
Posted by: Andrew | Nov 23, 2005 2:51:23 PM
Just from my observations, I believe the view that standardisation stifles creativity stems from the fact that proprietary extensions added to browsers are often seen as creativity from the browser vendors themselves, yet standards are seen if they're just getting in the way when in fact, it is the other way around.
This could be due a combination of factors, such as the standardisation process being relatively slow or because standards reject so many proprietary extensions, and are thus seen as rejecting creativity.
However, people many fail to see that proprietary extensions, while they may seem like innovative and flashy bells and whistles, end up doing nothing more than creating vendor-lock-in and, in fact, slowing progress.
For example, IE has so many proprietary extensions, yet standards like CSS have surpassed them all and IE is now one of the most significant, bug-ridden barriers preventing creativity and innovation, simply because we can't reliably use many superior features in the standards.
Posted by: Lachlan Hunt | Nov 29, 2005 12:09:17 AM
I come at this from a completely different perspective.
Many can measure the depth of their CSS/HTML experience in years. I measure mine in months. I am just learning web programming.
Many come from the design side. I come from the programming side. In the early 70's I started coding machine and assembly language on RCA Spectra 70 series mainframes. Writing early generation compiler code. I met and spoke with Grace Hopper a number of times. A delightful women. Don't ask. ;-)
"Graduated" to IBM/360 series COBOL. In the 80's I moved from programming to management and drifted away from the technical side.
Now I'm taking early retirement. Dabbling again. And nobody dabbles in the ancient languages I knew. I guess it's like speaking in tongues now. So I'm learning CSS, HTML and PHP.
Whatever. I'm amused by all this talk on standards and conventions. We went thru such revolutions and evolutions 30 years ago. Pundits like Ed Yourdon spouting about structured approach to coding and analysis. And resistance along the lines of it takes longer and gets in the way. And stifles my work.
And since I was teaching programming at the time, I ended up in the middle of it. My feeling remains the same. Structure and standards improves productivity. Short term and long term. It applies to the approach. The method. How you do it. Consistency doesn't get in the way. It helps.
Creativity is in where you want to get to. Creativity relates to the vision of what you want to do. What the end result should be. Not in how you get there.
Consistent methods and standards were applied to produce maintainable results in an effective manner. Regardless of whether we were producing a general ledger, payroll, warehouse management or MRP system.
The more things change, the more they stay the same.
Posted by: Walt | Nov 29, 2005 12:20:16 AM
Nice article, and a lot of good comments! I'm also supportive of standards, and agree they are not an obstruction for creativity.
I'm very much looking forward to your article on patterns. Web developers not seeing the point in using empirical knowledge and intuition for creating design patterns (which you could call (de facto) standards) should broaden their fields of knowledge. Design patterns and standards have become more and more important over the last decades as software engineering have evolved. I don't see why web developers should look upon themselves as an exception to this.
I'd guess Walt would be able to write more about that. I don't have half the experience he probably has :-)
Posted by: Frode Danielsen | Nov 29, 2005 12:51:28 AM
"IE has so many proprietary extensions, yet standards like CSS have surpassed them all and IE is now one of the most significant, bug-ridden barriers preventing creativity and innovation..."
On the other hand, IE introduced XmlHttpRequest, and a zillion creative, innovative AJAX applications followed. If we left it to W3C Web Standards, we'd still be waiting for this functionality.
There is a tension between standards and innovation. If every browser maker standardised HTML/CSS support by all releasing new browser versions tomorrow, we could do what we've been doing for years more easily/consistently. But we couldn't actually do any new things. The tags/functionality in HTML have not changed since 1997 - 8 years!
I want new things, even if they are "non-standard" to start off with, and it seems to me the browser makers are best-placed to innovate rather than the W3C. So how do we promote (and standardise) XmlHttpRequest-type innovations while avoiding proprietary lock-in?
BTW the Semantic Web uses RDF and OWL, not (X)HTML, so class/id's have nothing to do with the SW.
Posted by: mattur | Nov 29, 2005 12:56:11 AM
Frode - Experience that is 100% useless now. Not too many people looking for my expertise. But I'm guessing you know that. ;-)
Posted by: Walt | Nov 29, 2005 5:41:48 AM
Discussion of standardisation is often muddied by not defining what is meant by 'standardsisation'.
We're talking about (at least) two types of 'standards'. Conformance to W3C technical specifications and adopting recognised interface conventions.
The former is to code what spelling and grammar are to language; a formal shared framework.
The later goes more to notions of usability. A real world example being 'book' conventions. For example, numbers on the corner of a page (usually) identify the folio number. Familiarity with this convention enables us to find a page within a book.
If the location and visual treatment of an interface element such as navigation lists matches the user's expectations (or mental model) then there is less of a 'barrier' to use. A non-conventional navigation element will require the user to learn a new convention before they are able to interact with confidence.
There's also the issue of experience to consider. How many developers currently have the requisite skills to create pure HTML/CSS (standards compliant) designs? If only a handful of people speak Klingon is it any wonder there are no literary masterpieces published in that language?
Posted by: Andy Kirkwood | Dec 3, 2005 8:43:14 AM
An aside - I have looked up the standards for the word 'standardize' :-)
Posted by: generic viagra | Jan 6, 2010 12:47:39 AM
You have a very good blog that the main thing a lot of interesting and useful!
Posted by: Buy Generic Viagra | Jan 22, 2010 5:15:23 PM
My dog is easily guessed or all!
Posted by: Compra Viagra generico | Jan 22, 2010 5:19:09 PM
The argument is that standards are bad, because somehow standardization will stifle innovation and creativity. I frankly can't see how that logic works.
Posted by: generic cialis | Mar 12, 2010 6:43:38 AM