Of course, 140 characters couldn’t capture his whole opinion on the subject, so he wrote a blog post followup. I strongly recommend reading it. I really agree with it.

Also, he later tweeted: “My new opinion, based on insight from commenters: You DON’T need to know how to write HTML/CSS. But in many scenarios it can be useful.”

I agree with this too, for the reasons explained well by Mark Boulton. A web designer doesn’t necessarily need to know how to write actual HTML and CSS. But she does need to know about the limitations imposed by browsers, the strengths and weaknesses of the medium. Writing HTML/CSS can make understanding and designing around these limitations a lot easier. But you can come to this understanding without learning HTML/CSS first.

Basically, my opinion is still the same as it was when I wrote Should web designers build their own pages? back in June 2008. Check that article out for my full opinion—including my very own architect analogy! (Seemingly a must whenever debating this topic.)

While Andy’s proposal doesn’t just pertain to the design process for flexible sites—he seems to advocate it for all types of sites—Andy does ask:

Is the fact that so many web pages are fixed width and centered a direct result of clients signing off fixed width design visuals?

I’m willing to answer yes to that question. Not that this is the only reason fixed-width designs get created, but I’m sure it plays a role in many of them. As I talk a great deal about in my book, as well as the presentation I’ve made at a couple conferences lately, you can’t build a flexible page unless you’ve designed it to be flexible in the first place. You may design a page that must be fixed-width inadvertently, and won’t catch it until you go to build the page in (X)HTML and CSS. But if your client already signed off on the design when it was a static image, it’s now too late to change the design to make it liquid-compatible. Forcing yourself to build at least a rough version of the page before you present the design to the client is a great way of checking if your design is truly flexible-friendly. If you find that it isn’t, you can make design changes to avoid CSS headaches later.

A lot of the commentary surrounding Andy’s proposal has been that building pages for clients will take a lot more time that isn’t in the budget. I think this is a valid argument. But, I also agree with the arguments that this workflow method can actually be a time-saver—or at least that the extra time spent making the (X)HTML/CSS comp can be made up later. Andy says:

As I work almost as fast in markup and CSS as I do in Fireworks, I find that the overall time that I take designing and implementing that design is dramatically reduced as I am not duplicating work. It leaves me free to spend more time on fine tuning a design and I’m sure that my designs are all the better for that.

I’m right with Andy on this one. I’ll spend a lot of time futzing with the perfect border color or the perfect amount of spacing under a heading; when it comes time to build the page, I’ll have to do all that futzing over again in the CSS to find the perfect em measurement, for example, that matches the spacing I created in the Fireworks comp. It’s often faster for me to just leave all the work on these design details to my CSS development stage.

Another argument for time-saving comes from Kyle Weem’s response to Andy’s original post:

…you can absorb that extra time spent in the budget that would have been wasted on extravagant solutions generated to create identical rendering where it need not exist.

Another argument I totally agree with, especially if you are trying to create flexible designs. When you build the (X)HTML/CSS before showing to the client, you catch things in your design that are going to be tough and time-consuming to build—either because you’ve designed them in such a way that you can’t make things flexible, or it’s just going to take a lot of hacking to get it looking the same cross-browser.

All this said, this method of showing (X)HTML/CSS pages to clients instead of static design comps is not something I’ve actually done (though I do sometimes show wireframes before design comps). But as someone who primarily designs liquid layouts now, I can see it being really beneficial to the way I work, and I’m definitely going to give it a try soon.

My answer in a nutshell: Ideally, yes.

I’ve been lucky enough to always design the pages that I build. I’ve worked at a design agency, as part of in-house design teams, and freelance, but never have I had a separate designer create comps to then hand off to me to build.

I call myself lucky in this regard because even though I’ve never experienced this design process firsthand, I’ve seen how frustrating it can be for other web designers. (Note that I said “can,” not “is.”) When the visual designer doesn’t know how to actually build web pages, he or she frequently designs in elements that just don’t work well on the web. Either they’re things that cannot be built with HTML and CSS—such as elements that depend on an exact font size—or they’re things that can be done but are either difficult or unwise to implement, due to current best practices in the areas of standards, usability or accessibility.

It’s been my experience that the more experience a designer has actually building web pages, the more successful he or she is in designing interfaces and graphic elements that work well with its strengths and constraints. That doesn’t mean that every designer should build his or her own pages; it’s fine to have a division of labor, and I’m sure it’s actually the smartest process in many teams. Nor does it mean that the designer has to be an advanced HTML or CSS developer. It simply means that the designer ought to have some idea how to build the page if he or she really had to.

Mark Boulton argues in his post “Design isn’t about tools” that “knowing how to build things in HTML/CSS limits your creativity.” I see his point and think it’s a really good one. We all have to be careful to not just keep designing things the same tired way because we’re used to building things a certain way. It can be a really positive, fun experience to drop a print designer into your comfy design nest to force you to consider new design possibilities, even if you can’t use them exactly as-is on the web.

But at the same time, it’s unrealistic to think that you can design without the constraint of any tools. At the Voices That Matter conference I was just at, Jason Cranford Teague gave a presentation on web typography in which he said that design is about constraint—working within it as well as overcoming it—and that if you want to work without constraint you’re an artist, not a designer. I think this applies to the question of whether designers should feel constrained by HTML and CSS, and I think the answer is yes. Architects don’t have to be qualified to build houses, but they have to have some idea of what construction techniques are possible, the physical characteristics of the materials they’re designing with, general physics, etc. They can’t just design buildings without any knowledge whatsoever of how they will ultimately be built.

Mark uses a print design analogy to ask “do print designers need to know how to build a press?” This analogy makes no sense to me. It’s like asking if web designers need to know how to build a browser, or a computer even. Of course not. The real question is “do print designers need to know how a press works”—and the answer to that is an emphatic yes. They don’t need to be qualified to actually operate a press, but they need to know the limitations and strengths of the printing press they’re work is going to be produced by. Print production processes are a pretty important part of the curriculum of any good graphic design program. (That last sentence had a lot of p’s in it!) Print designers need to know practical things like not to design important info within 1/8″ of the edge of a piece, not to run something that must be matched up perfectly across the fold of a multi-page book, what type of paper they’re going to be printing on, etc. I suppose you could call this being constrained by your tools—but it doesn’t necessarily follow that that’s a bad thing.

(On a little tangent, I applaud Jon Hicks for pointing out during this whole debate that Fireworks is a better tool for creating web comps than Photoshop. I switched to Fireworks for web work almost five years ago and can’t imagine going back.)