I normally run my screen at 768p, 720p, and lower; so
I have a lot of disdain for sites which plastered its own (usually very thick)
header or footer as a fixture on my browser's viewport
via `position: fixed` CSS rule-- which I figured this was the "sticky footer" that OP was referring to.
Doing so
makes my view of the site a short window and
hamper readability as well as my ability to skim. Sometimes, the view would be so short to the point that I couldn't see a single paragraph/image in full, despite that my screen resolution was supposed to be enough to handle them.
When encountering such thing, I would flick the trigger of CSS-killer cannon, tearing off that content-obstructing peepwalls into pieces,
to continue reading in peace.
I would never inflict such thing on my own websites.
Note that unlike some other comment here, I see sticky header
(1) and footer equally as nuisances. My keyboard has perfectly-functional Home and End key, and I know where the navigation header and info footers are on the page; I can get to that by a single flick of finger if I wanted to. Dangling them in front of my face all the time while I was trying to read the actual substance between, wears out my patience fast.
For regular footers (which I figured that's what the OP referred to as "fixed" footer?) which would stay at the end of content even they weren't filling to the bottom of the screen:
I don't mind them at all, that's a just how HTML works;
and it's the way I normally do whenever I add footer to whatever pages I crafted.
However, for sake of adding something useful other than just stating my own opinion here: it might be slightly off-topic, but I would like to note that
there is "the third way" too:
Once upon a ye olde time of 2007, some dude came up with
a magic CSS sticky footer hack that keeps it staying at the end of scrolling content when the browser's viewport was too short to display them all-- like a regular footer; but once the browser viewport became tall enough to have space below the content, this footer would now magically stick to very bottom of the window-- like a "sticky footer", instead of dangling at the end of content.
It was done with no JavaS'creep, in CSS2-compatible
(2), browsers-portable, and
*gasp* IE6-ready way. The only catch is you must know (or prepare to allocate surplus of) the height --in EM, pixel or whatever-- that your footer is going to have in advance.
(3)Following link is the original demonstration and the code... but note that because Internet Archive's banner interferes with it,
you MUST disable JavaS'creep here for the thing to display correctly:
https://web.archive.org/web/20160206003022/http://ryanfait.com/sticky-footer/Rest in peace Mr. Fait-- the dude who came up with this great piece of style.
(1) Header plastered with `position: fixed` also bears another sin: it interferes with links to subsection or paragraph. When such header was in place, it covered up few (or several) topmost lines of the current viewport; which when such link was clicked, the browser would position the link's target as the top line of the viewport-- right behind that pesky obstruction, which caused mental double-take about what exactly that link had led me to; and required me to scroll back for several lines to actually see where the darn thing actually was. Nightmare for long and cross-reference heavy pages, and basically all pages with table of contents.
It requires another horde of workarounds to fix such links/anchors to work sufficiently again; so a number webmasters who inflicted such careless obstructions in the first place would sometimes just left them wonky like that, and I'll fire CSS-killing cannon whenever I realized such wonkiness were there.
(2) No media query, or CSS/MovingGoalpost was involved.
(3) When you use this code, no usability would be harmed even when your actual footer content overflown the allotted height. The only "imperfection" you would see in that scenario was when the browser's viewport height was high enough for both content + actual footer height to supposedly be entirely on-screen, the scrollbar would not disappear.
In practice, it would be unusual to actually see such imperfection in practice: because the most common reason of such overflow was low horizontal resolution causing text to break into more than expected number of lines. But screen with low horizontal resolution usually come with low vertical resolution to, which caused the content to scroll-- hiding away the imperfection in such case anyway.