what's the best i can expect for english text-only lossless
data-compression, in terms of ratio of raw data/compressed data, using
a current state-of-the-art public domain algorithm?

Hello Walton,

The "best" ratio of raw data (English text) to compressed data is
roughly a ratio of
  8 : 2.3
[or a single byte (8 bits) of text is encoded into 2.3 bits]
  http://www.cs.waikato.ac.nz/~singlis/ratios.html

For "popular" methods, the ratios are roughly
  8 : 4 Lempel-Ziv or (.zip)
  8 : 2.7 Gzip
  8 : 2.3 Bzip2
Source codes for each of these compression algorithms are available on
line. For example:
  http://www.gzip.org/
has links to both algorithms and source code for gzip.

Notice also that some algorithms are adjustable. For example, with
bzip2 there are several settings possible as described at
  http://www.mkssoftware.com/docs/man1/bzip2.1.asp
[scroll down for a table of results with the Calgary corpus]
The results will also vary based on the input file (size, type of text).

A few other references:
  http://en.wikipedia.org/wiki/Data_compression

Search phrases:
  lossless text compression ratio

  --Maniac

exactly the answer i wanted.  i especially appreciated the links for
additional info.

Not sure if its the best compression ratio but it is good.
http://www.7-zip.org/7z.html

rzip takes a bit longer, but uses a larger sample space resulting in
significantly smaller compressed files.  http://rzip.samba.org

RAR provides very good compression ratio for plain text files ( about 8:2 )
http://www.rarlab.com/