This is a fantastic post and discussion. I've had many a debate about this, and I tend to come down on Jeff Croft's side.

For the record, I'm more of a developer than a designer, but design was my first love, and I've been wrangling HTML by hand for 15 years now, and CSS for 10.

I see the value of Sass. CSS has some really obvious pain points when it comes to things like standardizing color palettes and certain types of design constructs. It doesn't take much to add a lot of value to CSS, which I think Sass does admirably.

But my problem really boils down to the ubiquity of HTML and CSS. Everyone knows it. Windows, Mac, .NET, LAMP, it doesn't matter, everyone uses it. All the tools support it, and all computers have a web browser where you can view the raw files directly.

As soon as you add a pre-processor like Sass that simplicity is gone. You can no longer work with any designer, you need one that has the pre-processor set up and is comfortable with the syntax. In a pinch you can pass the processed CSS to a designer, but then you have a versioning problem, and you are likely creating lots of busy work for yourself.

You face a similar issue with tools. They all support CSS, and some of them have some really cool features that can only come from having a userbase of a million developers battle testing it over years. It's not just text editors I'm talking about here, but debuggers, spiders, javascript libraries, and who knows what else.

So the real issue with any kind of CSS preprocessor is that by choosing it you are limiting your ability to leverage a whole ecosystem of talent and tools. Ideally a really great pre-processor could someday get enough traction to see itself become a standard, but that is not something we can hope for any time soon.

Is the tradeoff worth it?

For me I default to saying no, because I already have learned how to deal with CSS and it's an insignificant pain point to me compared to all the other things that go into building a great web product.

However in the general case costs vary. If your canonical design already lives in a framework like Rails then Sass doesn't really incur any additional tooling cost, because anybody working on the design already has to have the framework running locally anyway. However there are many team structures and workflows where HTML/CSS mockups ideally can stand alone.

I would love to join a team working with Sass, but I don't think I would choose it for a greenfield project (yet).