Context is important. A no-Photoshop approach makes sense for Quora since their product is text-heavy. For a product that requires, say, rich data visualizations like Google Analytics, some design ideas need to be mocked in a graphics package (Googlers often use Fireworks instead of Photoshop) first, then prototyped in code.
Context is also important in terms of what stage of the development process the designer is at. If a product design is already well defined (e.g. "top nav is global, 2 columns in the core content space, right column is contextual nav, left column is the content stream, here's the style guide ..."), then going directly to code makes sense. As Joel suggests, going directly to code is great for day to day new feature iterations.
But if your product is completely undefined, Photoshop or Fireworks can be excellent tools for mocking up design variations and sets of use cases or states within those variations.
Hats-off to Quora for going straight to code when it comes to design. It works for them but it's not going to work for everyone in every case.
I tend to think that whatever allows the designer to create and communicate their vision the fastest is what's best suited for the task. For many designers, its the tool they've been using for over a decade. It imports and exports a wide range of formats, so even the case against proprietary is a non-issue. And when the project moves from the mock-ups stage to the creation of the graphical assets for the launch, Photoshop is still good for that too!
Without question, Photoshop is the most versatile tool I've used in my professional career as a web developer.
>I never understood why people are buying proprietary, expensive stuff like Photoshop just for mock-ups.
People have been using Photoshop for years (like myself), and personally these people feel that open source equivalents like GIMP are still way off in terms of competing with Photoshop.
>First of all, vector graphics tools like Inkscape are a lot better suited for that task.
I have used both Inkscape and Photoshop for quite a few years. It's unfair to even compare them. And, frankly Photoshop _is_ far more powerful if you want to go for high fidelity mockups.
In my experience, Fireworks is far less useful when it comes to the creation of graphics after the initial wireframing. On the other hand, Photoshop [and GIMP] is completely usable throughout the entire design process. For this reason, many designers would rather learn, and pay for, only one product, and that product is Photoshop. As a side note, although GIMP lags in a few areas, notably on the typography side of things, it's the best product for my needs when I weigh cost and features, and only rarely do I find myself wishing I had tossed several hundred dollars for Photoshop.
>> Second, even if you don't like vector graphics, there are great Free Software projects like Gimp and Krita that deserve our attention and our support.
I'm not sure that I follow here. Why do they deserve our support? (Not to say that they don't deserve support, but thats a rather imposing claim without support)
> I never understood why people are buying proprietary, expensive stuff like Photoshop just for mock-ups.
When I was younger, I couldn't imagine why people would spend hundreds on Creative Suite, MS Office, etc. I could just use GIMP, Inkscape, OpenOffice, etc. for free, and stick it to the man.
Now, I run a business, and I think anyone who uses substandard tools to save a few hundred bucks is crazy. I also think anyone who really believes that the freebie tools I mentioned are even close to the standard of the professional tools I mentioned is crazy. The cost of these professional tools is relatively small compared to the money they save through increased efficiency alone, but in many cases they get much better results that would justify the cost anyway.
(I could list a numerous objective, widely applicable reasons in each case, and I have done so in other forums on several occasions, but this probably isn't the place. I'm just trying to explain why people do the things you asked about.)
Although your statement is apparently a joke, there's some unexpected truth in it:
You don't even need a hex editor – SVG is an XML format, so a simple text editor will do perfectly fine! This means that in case of some Inkscape issues, you can always fine-tune the vector graphics "by hand".
This also means that generating vector graphics is as easy as generating HTML code in a web application. You don't need any imaging library for that. (For the old-school hackers: yes, this is also possible with Postscript, i.e. EPS) Then, you can use Inkscape to preview the picture and to make some final changes by hand, if desired. (Here, things get tricky if you use EPS instead)
Also, any modern browser understands SVG, so you don't have to do any image conversion. Heck, you don't even need an <img> tag! Just put the SVG code directly into your (X)HTML document.
tl;dr: SVG is really cool, and Inkscape uses it natively.
Context is also important in terms of what stage of the development process the designer is at. If a product design is already well defined (e.g. "top nav is global, 2 columns in the core content space, right column is contextual nav, left column is the content stream, here's the style guide ..."), then going directly to code makes sense. As Joel suggests, going directly to code is great for day to day new feature iterations.
But if your product is completely undefined, Photoshop or Fireworks can be excellent tools for mocking up design variations and sets of use cases or states within those variations.
Hats-off to Quora for going straight to code when it comes to design. It works for them but it's not going to work for everyone in every case.