home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.apollo
- Path: sparky!uunet!nih-csl.dcrt.nih.gov!helix.nih.gov!rvenable
- From: rvenable@helix.nih.gov (Richard M. Venable)
- Subject: Re: Need Some Advice..
- Message-ID: <1992Nov11.090143.901@alw.nih.gov>
- Sender: postman@alw.nih.gov (AMDS Postmaster)
- Organization: National Institutes of Health, Bethesda
- References: <1992Nov5.163551.2615@ttsi.lonestar.org> <1992Nov5.201427@helix.nih.gov> <1992Nov9.192837.17737@bcars6a8.bnr.ca>
- Date: Wed, 11 Nov 1992 09:01:43 GMT
- Lines: 92
-
- In article <1992Nov9.192837.17737@bcars6a8.bnr.ca> djf@bnr.ca writes:
- >In article <1992Nov5.201427@helix.nih.gov>, rvenable@helix.nih.gov (Richard M. Venable) writes:
- >|> ...
- >|> token ring, we plan to run Domain/OS as long as possible. Frankly,
- >|> NFS,
- >|> ethernet, and X windows are a step backwards as far as many long time
- >|> Apollo
- >|> users are concerned.
- >|>
-
- The author of the above responds:
-
- >Not to start a religious war or anything but, as a user of both windowing
- >systems, I'm interested in knowing what people see in Domain pads, editor,
- >etc. Sure the ability to search back through the pads was nice, but the
- >functionality exists with Xterms - just not as gracefully.
-
- It's that grace and simplicity that some of my cohorts are stuck on. Myself and
- several others made the transition from vt100 on a mainframe or mini directly
- to Apollo workstations. Unix was an infant, regarded as an AT&T curiousity,
- and X-windows didn't exist. Compared to using vt100 and tek4010 terminals
- on VAXes and IBM 370s, though, the DM was very nice indeed. Some things we
- especially like are: (1) the ability to move the mouse cursor to a file name
- in a pad and open an independent pad to view or edit the file, (2) cutting and
- pasting rectangular blocks, and (3) "grabbing" blank delimited words from the
- output and appending it to the end of the input buffer. The DM editor, with
- it's additional left keypad, is very intuitive and easy learn; I can't really
- say the same for vi or emacs.
-
- >Other than that though there is little I miss. I found the window manager
- >really painful for reasons too numerous to mention, but most had to do
- >with how little I was able to customize it. There seemed to be too few
- >ways to do neat things that X, MWM, et al can do using resources, xset, etc.
-
- It's that grace and simplicity notion again; there aren't as many of those
- cute little clocks, mailboxes, and other Xtoys, partly because Apollo
- workstations are not nearly as widespread as Unix boxes with X. The DM is
- very efficient compared to X, and is especially well suited to our code
- development, scientific research, and molecular graphics environment.
-
- >Maybe I just didn't pick up the DM tricks along the way; if so I blame it
- >partly on the lack of documentation compared to O'Reilly and all the other
- >authors at various levels on X.
-
- I found the Apollo documentation, especially the on-line help, to be very
- helpful in learning how to manipulate and customize the display manager; much,
- much better than X and O'Reilly for the novice. Many people have trouble
- getting started with X because of it's complexity; but like any other window
- system, if there's a local expert available, that helps a lot.
-
- In spite of what I say here, I spend a good deal of my time convincing other
- Apollo users to switch to one of the Unix shells, and to use Xwindows at
- almost every opportunity. I do so for most of my own work, especially when
- accessing remote Unix hosts; it's crazy not to IMHO.
-
- >So what sorts of thing could DM do that X can't?
-
- Well, the DM could run efficiently on a node with a 60 MB disk and 2 MB of
- RAM; I don't know if X could do that-- but it's really moot. They are
- different approaches to controlling the display; but X is standard, Apollo
- workstations are no longer made, and the days of the Aegis shell are limited.
-
- However, HP has a DM emulator available for the 9000 series (model 7xx at
- least) principally because of people like us, who were loyal Apollo customers
- but can see no particular advantage to one Unix, X windows box over another,
- once the price/performance specs we want are matched.
-
- X was only one of the things I mentioned. Consider:
-
- - the token ring operates 20% faster than ethernet (12 Mb/s vs. 10 Mb/s).
-
- - the token ring degrades gracefully under load; no contention overload as
- with ethernet
-
- - one or two simple commands integrate a new node into the net; no tables
- to maintain; ALL volumes accessible to all nodes on the token ring
- (privileges permitting, of course)
-
- - little delay in accessing volumes on the token ring net; much more efficient
- than NFS
-
- Overall, the combination of the DM, the token ring, the Apollo // network root
- model provided an efficient and elegant environment for our activities, and
- still does.
-
-
- <---------------------------------------------------------------------->
- < Rick Venable (O O) | "Back off, I'm a scientist." >
- < FDA/CBER Biophysics Lab )|( | -- Bill Murray in >
- < rvenable@helix.nih.gov (=) | 'Ghostbusters' >
- <---------------------------------------------------------------------->
-
-