[comp.sys.apollo] /sys/dm/color_map

gjalt@eutesap13.euteal.uucp (Gjalt de Jong) (04/07/89)

If you change the color table of your node, by modifying the file
/sys/dm/color_map, and reloading it with the 'lcm' command, the color
table of all the partner nodes are also changed. 

Wouldn't it be better if this file were a link to the `node_data directory?
Since it is, at least to me, node specific data.

BTW, we're still running SR9.7, so I don't know if it's changed in SR10.x.

--
Gjalt G. de Jong,                 | Phone: +(31)40-473345
Eindhoven University of Technology, Dept. of Electr. Eng.
P.O. Box 513, 5600 MB Eindhoven, The Netherlands
Email: gjalt@euteal       UUCP: {...}!mcvax!hp4nl!euteal!gjalt

oj@apollo.COM (Ellis Oliver Jones) (04/19/89)

In article <GJALT.89Apr7091124@eutesap13.euteal.uucp> gjalt@eutesap13.euteal.uucp (Gjalt de Jong) writes:
>If you change the color table of your node...
>with the 'lcm' command, the color
>table of all the partner nodes are also changed. 
>Wouldn't it be better if this file were a link to the `node_data directory?

Right you are.  It would be better.

>...I don't know if it's changed in SR10.x.

It's changed in sr10.1.p for the DN10000 (Real Soon Now) and
will be changed in sr10.2 .

/oj (speaking for myself, not for Apollo Computer, Inc.)

deborah@CITI.UMICH.EDU (Deborah Swanberg) (04/21/89)

	
	In article <42b70de8.d5b2@apollo.COM> oj@apollo.com (Ollie Jones) writes:
	>In article <GJALT.89Apr7091124@eutesap13.euteal.uucp> gjalt@eutesap13.euteal.uucp (Gjalt de Jong) writes:
	>>Wouldn't it be better if this file were a link to the `node_data directory?
	>
	>Right you are.  It would be better.
	>
	>>...I don't know if it's changed in SR10.x.
	>
	>It's changed in sr10.1.p for the DN10000 (Real Soon Now) and
	>will be changed in sr10.2 .
	>
	>/oj (speaking for myself, not for Apollo Computer, Inc.)
	 
	Why don't you just make it a link.  I did along time ago.
	There are several files that would be better of being links
	that I have made into links without problems.  Just use your
	preferred method of creating a link, either AEGIS or UNIX.
	 
	Mike Markley
	Apollo Systems Support 
	UCSD Department of Electrical and Computer Engineering
	 
	 
	I always speak for myself.  No one listens that way.
	
	
No, I think this is still wrong because it works only in an environment
where an Apollo is a personal workstation assigned to one person. 
In an envoronment where many people share the worstations, each person
may want their own color scheme, and the original dictation of the
sole color map is revisited.  The lcm command allows a color map
at any location to be loaded.  I have a directory of color maps I've
collected from various people, load my favorite from my user_data startup file,
and change it according to the amount of light in the room.


Deborah Swanberg

Center for Information Technology Integration (CITI)
University of Michigan
2901 Hubard
Ann Arbor, MI  48105
313-763-7479

deborah@citi.umich.edu

FERGUSON@TMASL.EXXON.COM (04/22/89)

Why not just make the link to people's user_data directories, and
copy it in to all active accounts then? You can put these files anywhere
you like, just remember that Apollo software updates don't assume
anything and you'll need to re-create your changes with every update.

Scott
ferguson@erevax.bitnet

oj@apollo.COM (Ellis Oliver Jones) (04/22/89)

In article <GJALT.89Apr7091124@eutesap13.euteal.uucp> gjalt@eutesap13.euteal.uucp (Gjalt de Jong) writes:
>Why don't you just make it a link.  I did along time ago.

Yes, so did I on my own node.  However, read the excerpted release notes included
at the end of this article to see a problem that isn't solved by
a simple `node_data link.
***

In article <8904201526.AA04468@umix.cc.umich.edu> deborah@CITI.UMICH.EDU (Deborah Swanberg) writes:
> ...The lcm command allows a color map
>at any location to be loaded.  I have a directory of color maps I've
>collected from various people, load my favorite from my user_data startup file
>and change it according to...

This is a fine idea.  Are you asking us to change something?
It doesn't seem like we need to change anything further so you
can continue to do this.
***

Here's an excerpt from the release notes going out on SR10.1.p and
SR10.2 describing how the new scheme works.  (caution; don't
use stuff you read on news as a reliable source for release notes,
but rather read the notes that come with your software, because this
stuff could still be changed at the last second).

"At SR10.1.p we are moving the location of the system color maps in
 order to allow diskless nodes to have different color maps than their
 paging partners.

"Prior to SR10.1.p, the system color map was stored in
 /sys/dm/color_map.  At SR10.1.p  color_map is
 stored in two directories:

       'node_data/etc/dm_display/color_map
 and
       /etc/templates/dm_display/color_map

"Note that /sys/dm/color_map remains as a link
 to the color map in the 'node_data directory."

"When a diskless node boots from a paging partner that it has not used
 before, the system software copies the color map in
 /etc/templates/dm_display/color_map into
 'node_data/etc/dm_display/color_map.

"You can then modify the color map for the diskless node by editing
 'node_data/etc/dm_display/color_map.

"Note that if the diskless node has used the paging partner before the
 system leaves the color map file alone.

/Ollie Jones  (speaking for myself, not necessarily for Apollo Computer, Inc.)

freedman@cpsc.ucalgary.ca (Dan Freedman) (04/23/89)

In article <8904211745.AA10904@umix.cc.umich.edu> FERGUSON@TMASL.EXXON.COM writes:
>
>Why not just make the link to people's user_data directories, and
>copy it in to all active accounts then? You can put these files anywhere
>you like, just remember that Apollo software updates don't assume
>anything and you'll need to re-create your changes with every update.


I think you just answered your own question.  Having to re-create
things when the software changes is quite time-consuming.  What makes
it worse, is that it is difficult to tell ahead of time which changes
will need to be re-done after the installation.  The solution is to
have the vendor *assume* that people are going to want to customize
the o/s on both a system-wide and on a per-user level.  The vendor
then provides oodles of hooks that both system administrators and
users can hook their customizations into.  This is done already for
some things.  For example, both a system administrator and a user can
hook extra directories into the shell's search paths.  This kind of
flexibility should be extended to all system software.  True, we can
make our modifications and patches, but really, the o/s vendor should
make the software flexible enough so that user's own customizations
can fit in without changes to the o/s.

Dan Freedman
University of Calgary Computer Science Department
2500 University Drive N.W.			      freedman@cpsc.UCalgary.CA
Calgary, Alberta, T2N 1N4	                   ...!alberta!calgary!freedman

dennis@peanuts.nosc.mil (Dennis Cottel) (04/26/89)

Dan Freedman says:

> The solution is to have the vendor *assume* that people are going to
> want to customize the o/s on both a system-wide and on a per-user
> level.  The vendor then provides oodles of hooks that both system
> administrators and users can hook their customizations into.

Exactly.  Another example: make the HELP command search a list of
directories.  A simple change to the program by the vendor, and lots
of ugly configuration tricks could be done away with.

Yet another example: the DPCC installation requires configuration
files to be modified down in /sys/dpcc making them global to the node
and not variable for specific users.  (Unless you make your own links,
but we're discussing a better way.)

Ok, now that I'm worked up: *all* the configuration files for various
products ought to be segregated into a common specific area.  That
way, rebuilding the disk becomes simply reinstalling the product
(binaries and unchangeable data files), and then reloading a single
small backup of this node's specific configuration directory.

	Dennis Cottel  Naval Ocean Systems Center, San Diego, CA  92152
	(619) 553-1645      dennis@nosc.MIL      sdcsvax!noscvax!dennis