cleanup: KATANA in emulator C source is MAIKO in Lisp source

[nbriggs]

When FX was developing the original emulator their machine type for medley running on the emulator was Katana (sword?). I don't know the history of why the emulator became Maiko (apprentice geisha?), but in the C source we find:

% grep -ri katana inc src include
inc/hdw_conf.h:#define KATANA		3
src/allocmds.c:                        This version works only for KATANA-SUN
src/dspsubrs.c: *			called from  \Katana.DisplayWidth.
src/dspsubrs.c: *			called from  \Katana.DisplayHeight.
src/initsout.c:  InterfacePage->machinetype = KATANA; /* 3 is katana */
src/storage.c:  if ((*MACHINETYPE_word & 0xffff) == KATANA) {
src/storage.c:    error("set_storage_state: Sorry! We can work on only KATANA");
src/subr0374.c:                subr----0130 for katana trace
src/vmemsave.c:  InterfacePage->machinetype = KATANA;

while in the Lisp code, references to machinetype

                              (COND ((EQ \MACHINETYPE \MAIKO)
                                     (MAKERECORD 'IFPAGE MAIKO.IFPAGELAYOUT))
                                    (T (MAKERECORD 'IFPAGE IFPAGELAYOUT)))

and

                          (* ; "Machine types")
                          (\MAIKO 3)
                          (\DOLPHIN 4)
                          (\DORADO 5)
                          (\DANDELION 6)
                          (\DAYBREAK 10Q)

If we clean this up, it should probably move to Maiko.

[masinter]

Oh the emails

Date: 18 Dec 87 16:10 PST
From: Bane.pa
Subject: Minutes from emulator meeting 12-18
To: Shih,Snow,Tateno,Masinter,Charnley,Wozencraft,Sybalsky,Krivacic,BVrolyk
cc: Bane.pa

The meetings seem to be growing: 8 in attendance this time.

Action Items reports:

Frank Shih:

The emulator sources will live on {POGO:} and {POGO:}. Mail will be archived in {POGO:}MAIL>.

Bob Bane:

The official Webster people who handle Sun XNS software haven't replied to my message asking for it yet. I'm gone on vacation until January 5.

Masa Tateno:

Fuji requests some way to participate in our discussions. Perhaps video taping them? Larry suggested recording everything in mail and copying them on that.

Fuji suggests an alternate name for the project of MAIKO, which means "apprentice dancing girl," more or less. We agree; the DL for this project will be named appropriately.

Fuji can send people here for 6 months or less on an easy-to-get visa; longer than that will be harder. In any event XAIS needs to ask for them before they can come.

Don Charnley:

Can understand the C code for the emulator easily, but need to come up to speed on the rest of Unix. Larry suggested Don review the sources looking for things that were awkward in the microcode due to underlying hardware constraints and were carried into the emulator.

Larry Masinter:

Made a loadup with NoteCards and a large notefile in {CORE}, which crashed. Might have been caused by many things (wrong emulator, bad loadup procedure, emulator bugs); need to find out which one.

PS Smalltalk old documentation and sources are available to us.

New Stuff:

Frank Shih:

Talked with Walt Samuelson to get Marketing's view of some of the emulator issues:

Compatibility: source compatibility is probably sufficient, compiled code compatibility would be a plus.
Handling different resolutions: just displaying stuff smaller is OK since other Sun software does that.
Which Sun: Aim at Sun 3 and performance tune for it without precluding Sun 4 later.
Which lisp: Aim for Lyric if we're going to finish by end of March, otherwise Motown.
File system issues: aim for Unix tool compatibility (EOL conventions, etc.).

Larry said that we should poll customers on these issues and not rely on Marketing's technical expertise. In particular both he and Beau Vrolyk think a Koto emulator might be important, and Larry thinks one might be easy to do from a Lyric emulator.

John Sybalsky:

Bob Krivacic's contract is signed, finally.

Porting emulator to other machines:

Consensus was that once the emulator runs on Suns, porting it to other Unix-based machines (386, microVAX) should be straightforward. Masa presented an analysis of potential money in porting: if it's worth doing once, there's easy money in doing it more than once since most of effort goes into first one. We should look into getting consulting wizards for target machines to scope out difficulties.

Pricing:

Parc Place Systems charges $3000-$4000 for its VM on workstations, $700-$1000 on Macintosh. The emulator priced at these levels looks somewhat profitable; we anticipate selling ~$1 million worth, which will just about pay for the anticipated development resources.

Next meeting for those who will be around: Monday December 28.

[Kirtai]

Just noticed that this issue wasn't closed despite the merge.

[nbriggs]

Thanks for pointing that out.