Re: lynx-dev reading sjis docs [was Re: lynxcgi problem]

[Henry Nelson]

> But I can't understand the individual functions which should be 
> inside of Lynx or outside of it.
> Are you saying processing 1-byte kana should be outside of Lynx?
> Is this right?

Not the "processing 1-byte kana" per se, the "manual override" feature
(using ^L to manually change the document character set in order to get
better performance [means Lynx already can do "good" conversion if it
can properly determine the document character set?]) in its entirety.

I (my single opinion, and not necessarily agreed upon by others, certainly
not by Hiroyuki Senshu) think that Lynx should do the best job it can to
convert the encodings with the information it can obtain.  That includes
1-byte kana processing, and all CJK for that matter.  *If* Lynx fails to
convert them properly, i.e., you get "mojibake," *then* the user should
use tools outside of Lynx, IMO.

Do not misunderstand that to mean that I think Hiroyuki's work was fruitless.
I think for CJK users on a _Windows_ platform, there *may* have been no other
choice.  Still, I think that, for example, file://localhost URLs and dired
functions could have been set to sjis from the beginning for Windows users.
Not being a Windows user (of Lynx anyway), I don't know how hard it would be
to include a proxy with the Lynx binary.  Does lynxcgi, the sample solution
I offered, work on Windows?  I simply do not know or understand problems that
Windows users might face.

__Henry