[FreeNX-kNX] freenx is dead !

[Kurt Pfeifle]

Kurt Pfeifle wrote:
> Simone Piccardi wrote:
>> Fabian Franz wrote:
>>> - Copy Sources to /usr/NX/bin
>>>
>>> $ cp nx* /usr/NX/bin/
>>>
>>> - Copy node.conf to /usr/NX/etc
>>>
>>> $ cp node.conf.sample /usr/NX/etc/node.conf
>>>
>>> - Run nxsetup
>>>
>>> /usr/NX/bin/nxsetup --install --setup-nomachine-key
>>>
>>> - Done!
>>>
>> Hi,
>> yes, these instructions are quite clear, but I do not understand why it
>> is necessary to do a violation of the filesystem hierarchy standard and
>> use a separate /usr/NX/ directory for everything (also for the
>> configuration file, that should be in /etc).
>>
>> Just curiosity, because I cannot see any reason to do this.
>>
>> Regards
>> Simone
> 
> One-word-answer: legacy.
> 
> 
> Many-words-answer:
> 
> Not so long ago, X11 itself was "in violation" of the filesystem hier-
> archy standard, and installed everything into the /usr/X11/ directory.
> 
> As you may be aware, NX is in a way a fork from X11 (to a small degree,
> to a much larger degree it is an extension of X11), and it is designed
> to be installed and used without conflicting with the original X11, but
> run side-by-side with it.
> 
> When the NoMachine developers started their long, arduous journey of
> spadework to develop the NX core libraries (something a lot of people do
> not know, tend to forget or willfully ignore), they obviously didn't
> want to start with doing the least important but most boring work first:
> re-arrange the original X11 sources to comply with some obscure not-yet-
> fully adopted FHS spec of the time (and where X11 itself still didn't do
> it either), *before* even adding a single line of their own code.
> 
> If anybody seriously expects that, he's an armchair theoretician whose
> opinion can't be considered to be much relevant (boy, can I be polite
> these days....)
> 
> The initial NX development took several years, and all that time the No-
> Machine developers kept tracking many changes and improvements to X11.
> 
> It was the easiest and most efficient way to get a working NX code base
> by using a cloned X11 source directory tree as the base to start (and
> keep going).
> 
> Making a forked X11-tree to host the NX extensions development just to
> comply with FHS would have been a f***ing waste of time, especially be-
> cause this would have been required to be re-done again and again, when-
> ever the developers upgraded to a newer X11 version.
> 
> Once NX was beyond a proof of concept, and had their first working im-
> plementation running, and even released to the public (that was a few
> years *after* they started their work) of course, that was still not
> the time when a transformation to FHS would create any advantage -- not
> for end users, nor for NX developers.
> 
> Meanwhile X11 has moved and became the FHS-compliant software now known
> as 'Xorg', but NX still is based off an older X11 version that was
> hosted at /usr/X11/.
> 
> Just like initially it was too much work with no advantage to move NX to
> FHS when X11 itself was still in /usr/X11/ -- in the near future it very
> likely will be too much work with not much advantages to move NX back to
> /usr/NX/ once they start using the newer FHS'ed Xorg.
> 
> ----
> 
> I hope you can now afford to stop caring about the rather unimportant
> sideline issues around NX-based software....


I've been made aware of some even more important (and good) reasons for
not following the FHS when installing NX. More important reasons than
the ones I gave above. Read about them here:

   http://www.nomachine.com/ar/view.php?ar_id=AR01C00130

I can understand these. They are even "heavy" enough to keep violating
the FHS even *after* the GPL source base of NX have moved to a newer
Xorg's FHS template layout (contrary to what I stated as "very likely
will be too much work").

-- 
Kurt Pfeifle
System & Network Printing Consultant ---- Linux/Unix/Windows/Samba/CUPS
Infotec Deutschland GmbH  .....................  Hedelfinger Strasse 58
A RICOH Company  ...........................  D-70327 Stuttgart/Germany