I looked up pgs. 434438 of the TeXbook. It appears that
what is in the Manual is not what I thought I had done.
I intended to do all of Lists:
1, 2, 3, 4, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15.
I think that there may have been a problem with the missing symbols, but this was a long time ago.
The missing symbols from these lists should be included if
possible.
But, adding nonTeX symbols is not, I think, a good idea.
Some thought could go into a mechanism for users to create
their own symbols.

I have had a look what symbols TeX supports [1] and are not currently implemented in Octave [2]. From those I've picked the ones I personally have seen in scientific plots and consider them useful myself. So here is my suggested (and maybe very personal) list:
[1] http://www.math.union.edu/~dpvc/jsmath/symbols/welcome.html
This webpage claims to list „The symbols listed in Appendix F of The TeXbook“
[2] https://www.gnu.org/software/octave/doc/interpreter/UseoftheinterpreterProperty.html
Here is the full list of TeX symbols in [1] which are not yet in Octave [2] (leftout stuff is marked with square brackets):

If more symbols are added it would be good to make clear
that the current set is taken from the TeX book. Additional
symbols have been added for some specific reason.
In any case the result should be able to be processed by
TeX since this is sometimes done to produce results
suitable for publication, for instance.

Given that we like to view Octave as a superset of Matlab, I don't have an issue with adding some more symbols to the list. But, I don't want to go overboard and add 1,000 new entries.
Would someone be willing to look at TeX and the Octave manual and come up with a list of additional symbols that would be useful?

Updating bug priority and severity. And of course there is always full Unicode support as a longterm goal, which if/when done might obviate the need to support more named TeX math symbols.

Table 15.1 is actually derived from the TeX Book.
I think that it is a good idea to adher to what is specified
by Knuth in his documentation.
There are other means of introducing TeX source code,
as explained in the Manual.
This was discussed about a year ago when Table 15.1 and
the code were updated to match the tables in the TeX Book.

We could change this into a wish list item, but currently Octave is only trying to match Matlab's character set. It's not terrifically hard, but it does take some work to add a new character to the TeX interpreter so if we're mostly happy with the ones we have then I would make this a low priority item.

Oh, sorry. Then I was mislead by the information in the mentioned book of jwe on "Octave 3" (link in my original post). Maybe that's a bit outdated now.
If those symbols currently are not mentioned in the Octave manual nor support by Matlab, then I personally wouldn't consider this to be a bug at all.
But now it's up to the TeX experts...

Thanks for your bug report. I will have to let someone with more expertise on TeX support comment on whether or not this will be fixed. Setting status to "need info" until that is resolved.
However, the set of TeX symbols that Octave supports is explicitly intended to match the set that Matlab supports. Here is the Matlab set:
http://www.mathworks.com/help/matlab/ref/text_props.html#String
and here is the Octave set:
http://www.gnu.org/software/octave/doc/interpreter/UseoftheinterpreterProperty.html
As you can see, the symbols you say are not working are not listed in either of these tables, so they are currently not expected to work. The only question is, should they be supported or do we strictly support only the same set of characters as Matlab?

Here is some sample script to show what I mean:
If I run this on Octave 3.8.1 as well as on Octave 3.8.2rc1 under Ubuntu Linux, then the TeX symbols in s and t are NOT shown on screen. This is using FLTK.
If I switch to Gnuplot, then most of those symbols appear on screen, just the one in t still does not appear.
I assumed that all of those symbols should work because of this reference: http://www.networktheory.co.uk/docs/octave3/octave_172.html
I don't know if this happens as well on Windows installations, but I would assume yes.
