HTTP compression might be a good idea, or an unacceptable risk depending on your situation.
Most modern browswers support this, and those who don't will get uncompressed versions of the content.
However, there is a small group of browser versions who have bugs. Those who claim to understand HTTP compression but that in reality don't, or don't completely or correctly.
This group includes old Netscapes, MSIE6 before SP1 with Realplayer installed, old versions of Mozillas browser engine Gecko, and most likely some others too.
Nobody should use these browsers nowadays, but if they do they won't be able to access the Moodle, or will have serious problems doing so.
It all depends on how important you consider maximum
accessibility, and what kind of contact you have with your users.
For HTTP compression in
Apache i would recommend using mod_deflate (included by default) together with a conservative configuration enabling it, for example:
# Insert filter
SetOutputFilter DEFLATE
# Netscape 3.x is rumored to have unspecified problems
BrowserMatch ^Mozilla/3 gzip-only-text/html
# Netscape 4.x is rumored to have unspecified problems
BrowserMatch ^Mozilla/4 gzip-only-text/html
# Netscape 4.06-4.08 is rumored to have some severe unspecified
# problems
BrowserMatch ^Mozilla/4\.0[678] no-gzip
# MSIE and browsers posing as it might handle it, there is
# however rumors about another bug than the one below
BrowserMatch \bMSIE !no-gzip !gzip-only-text/html
# MSIE 6 prior to SP1 (Jan 2002) might not handle it
#
http://support.microsoft.com/default.aspx?scid=kb;en-us;Q312496
# Unfortunately there does not seem to be a way to differentiate
# between different MSIE 6.0 versions.
#
# BrowserMatch \bMSIE.6\.0 no-gzip
# Opera is rumored not to have had any serious problems handling
# compression
BrowserMatch \bOpera !no-gzip !gzip-only-text/html
# Konqueror has had bugs related to compression, but as far as i
# know none that prevented accessibility or readability of
# responses with Content-Encoding: gzip
BrowserMatch \bKHTML !no-gzip !gzip-only-text/html
# Old versions of Gecko have had problems with compression. I
# have not found a later confirmed bug than this one:
#
https://bugzilla.mozilla.org/show_bug.cgi?id=deflate
# A crude way to filter Gecko versions based on the fix date of
# the bug:
BrowserMatch \bGecko/199 no-gzip
BrowserMatch \bGecko/200[01234] no-gzip
# Adobes PDF reader plugin has been known not to handle compression.
# Also PDF files usually contain compressed data.
SetEnvIfNoCase Request_URI \.pdf$ no-gzip dont-vary
# Don't compress already compressed file formats
# Assuming that .exe means a windows self extracting file.
SetEnvIfNoCase Request_URI \.(?:t?gz|zip|bz2|sit|rar|exe)$ no-gzip dont-vary
# This includes compressed image formats
SetEnvIfNoCase Request_URI \.(?:gif|jpe?g|png)$ no-gzip dont-vary
# Make sure proxies don't deliver the wrong content
# (Proxies might convert images between formats otherwise)
Header append Vary User-Agent env=!dont-vary
Passive gzipping is inappropriate for sites where nearly all content is generated by scripts, like in Moodle.