Separating form and content is like separating language and meaning—possible, but artful. It is artful because it cannot be neutral. It cannot be neutral because content can neither be represented without any form at all, nor presented without some influence on how it is read.
In webdesign plaintext is often treated as nonformat, but it is format too—it raises the questions of character sets, markup versus markdown, hard versus soft wraps—it is haunted by carriage returns and line feeds—and displaying it raises the question of monospace or proportional fonts, wrapping or truncating, syntax highlighting, &c. Nonprogramming designers can think of it as nonformat only because they do not have to deal with it directly, just as printers could think of handwriting as nonformat, because it was the stage in the process of printing that they had no responsibility for. But questions of form in plaintext are serious and important for those who have to deal with it directly. (If you think webdesigners are overearnest about grids, you have yet to ask an programmer about whitespace.)
Plaintext permanently enacted what typewriters had already established in practice—the idea of punctuation and capitalization as part of content, not form. In manuscripts punctuation was so often indistinct, and its use so inconsistent between writers, that printers were free to omit, add, or switch points as they pleased. Yet even though writers bred to the age of WYSIWYG use boldface and italics significantly—also underlining, which was once just how writers and typists denoted italics to printers—still, because bold and italic characters are not part of ASCII (the plain text standard set by typists) they are so often unavailable, inconsistently displayed, or subject to sudden disappearance between programs, that they cannot quite gain the status of content.
Of course the phrase—separation of form and content—usually frames as a goal what might be more blatantly put by webdesigners as a command: programmers shall not meddle with design—or put by programmers as a judgment: just get it working, you can figure out the design later.
The phrase is also misleadingly abstract. As used it is meaningful only within a certain scheme: a database contains plaintext which is then filled into a template and served to a browser. This is not the only way it could be done. Quite different architectures are possible. Consider RSS, for example, where there is no browser and no design.
RSS is much closer to the intentions of the web than the web itself is. The fact that the web has any place for designers at all is the result of its abuse. Originally the web consisted of absolutely semantic markup. All pages shared one design—the design they inherited from the operating system being used to view them.
The early dominance of Netscape, however, by rendering pages in a consistent way across all platforms, let us get away with sneaking in the habits of graphic design on paper—the very metaphor of pages suborned it. Then the war between Netscape and Internet Explorer was fought around incompatible features for visual presentation.
Eventually all this was reeled back into CSS, but that was a capitulation, not a victory: the idea that the web meant the freedom to design was irreversible.
In current terms, the web was meant to be like Facebook—you put information in and let presentation be handled upstream—but it turned out like Myspace. To call this system separation of form and content is a little disingenuous.
As for the second kind of neutrality: I will use the Ruricolist as an example. Superficially it may appear that the Ruricolist is a blog by accident. The content seems separable from the form. That is true: it is practical to abstract these essays and present them in those other forms; indeed I have printed them in books. Yet the content of the Ruricolist exists because of its form. Only because there exists a form so familiar and flexible as the blog did I feel free to try something unusual.
My object in the design of this blog has always been to keep it as transparent as possible. I chose Blogger as least distinctive of all blog services, Minima as the least distinctive of all blog layouts, and Times New Roman as (to me) the most transparent of all fonts. (True, Georgia has become the look of blogging almost as resistlessly as Computer Modern Roman has become the look of science. But I still notice it; I do not notice Times New Roman at all.)
I have made adjustments to Minima gradually, to make it plainer and more readable (to my easily tired eyes): putting the columns in the golden ratio, opening up the linespacing, imposing an approximation to the traditional typographic scale. And now that Minima is no longer inevitable—I never see new blogs use it—I have replaced the header and markedly increased the font size. I might abandon it altogether if one of Blogger's new layouts has so low a figurative index of refraction.
I could have run the Ruricolist off its own domain; but I chose not to, because independence would mean irrelevance. Here among millions, being unusual just makes me eccentric; off on my own, being unusual would just make me weird. And once I decided on Blogger I was tempted by fancy layouts. With unassuming Minima, I can get away with using semicolons and subclauses, saying withal and writing sovran, and have it seem I refuse to be forbidden any of the resources of the language, I refuse to write down—not that I regret the century I was born in.
Presenting the Ruricolist with, say, Scribe would turn it into an exercise in antiquarianism. Likely it would be more popular on those terms, but they hold no interest for me. Writing an eighteenth century essay series as a blog is the kind of idea that, although clever in itself, gains nothing from actually being done. I wanted to extrapolate the form, not pastiche it.