[comp.sys.sun] Need information on SunView 2.0 and compatibility

levinel@uunet.uu.net (Lewis Levine) (12/13/88)

For the longest time we've been hearing how SunView 2.0 is going to
revolutionize Sun's display methodology. My problem with this is how am I
going to live with this change? Sun's track record seems to be fairly bad
lately in the backward compatibility area. If anyone can outline what
SunView 2.0 will support in both the new features and compatibility
department, I would be greatfull.

Thanx in advance.

Lewis Levine
Teradyne, Boston.
Phone: (617) 482-2706 x3294
teradyne!levinel

chuck@trantor.harris-atd.com (Chuck Musciano) (12/21/88)

> If anyone can outline what
> SunView 2.0 will support in both the new features and compatibility
> department, I would be greatfull.

Well, I went right to the source at SUG (Tony Hoeber) and asked how hard
it would be to convert to View 2.  The first answer is, "A mostly
mechanical operation".  When pressed, I discovered that all of the gadget
sizes change, so I will be forced to recompose every panel in my tools.
When I complained about this, Hoeber really didn't have an answer.  Now,
the question is, do I want to do the work to become View 2 (real name:
Open Look) compatible?  Does Open Look give me so much added functionality
that I simply cannot avoid it?  Right now, I think the answer is "no".

Unfortunately, I think that Sun View applications will not run under the
X11/NeWS server without being recompiled as View 2 applications.  I am
seriously considering creating an X toolkit that is identical to Sun View,
so I can seamlessly port my tools to X11/NeWS.  Anyone out there have such
a beast?  Anyone interested in working on one?  How hard could it be?

In an earlier posting, I asked for people's opinions regarding Open Look.
I'll start the ball rolling with these:

     1) In the December '88 Byte, Hoeber says: "The Open Look design
	team envisioned a typical user who wants to switch easily between
	Open Look, the Mac Finder, and the Presentation Manager.  We
	therefore ruled out design possibilities that would make this
	switch too difficult".  Show of hands please: how many Sun Spots
	readers find themselves making this switch?  I may rarely use the
	Mac, and I by God will never use the Presentation Manager, but
	I'll use this Sun every day for (hopefully) many years.  I think
	Sun users are being sold out for mass market appeal.

     2) Several times during the SUG Open Look seminar, when asked why a
	certain feature was or wasn't included, the answer was based upon
	"psychological reasons" or attempts to "reduce mouse motion".  So,
	I asked Hoeber exactly what cognitive psychological analysis Sun
	had done to measure the interface.  The reply: none.  Hoeber says
	that to do so would have meant that Open Look would have been
	tuned against only one criteria, rather than several criteria.  I
	contend that psychological testing of what is being pushed as "The
	Interface" is warranted, and should be considered as one of many
	evaluative tools.  In light of the fact that no testing was
	performed, Hoeber et al should stop using psychological reasons
	for feature justification.  The bottom line is that features were
	included based upon the intuition of the Open Look team.

     3) I have real problems with the scroll bars.  The elevators are
	small, and (I think) require some serious hand/eye coordination to
	hit the correct button.  Visual distinction between the elevator
	cable and the solid cable (indicating the size of the displayed
	region) is hard, much harder than the scroll bars currently in use
	in Sun View.  There are not specialized mousing events for scroll
	bars like Sun View which allow the various "jump line to top" or
	"jump top to middle".  This was done to make the interface easier
	to learn.  Again, I think this is a tilt towards novices at the
	expense of experts.

     4) One nice feature is pinup menus, which hang around, letting you
	select multiple buttons from the menu.  However, there is no clear
	consistent way to determine which pinup menu is associated with
	which application.  Hoeber indicated that color was one way, but I
	have a monochrome system.  A problem with no apparent solution.

     5) The raging controversy over click-to-type versus mouse-focus makes
	me wonder exactly who was interviewed to design Open Look.  The
	design team seems truly surprised by how adamant people are about
	the style they like, yet thay also tell us that many people were
	considered before Open Look was completed.  The issue of
	click-to-type never came up?  If click-to-type is not a user
	configurable option (which the Open Look team was considering) I
	will NEVER user Open Look.

     6) Another undefined issue (at least, my office mate could not get
	the answer from an Open Look person): if you have click-to-type
	set, and click in a window, does this also set the insert point in
	text windows?  If so, isn't this confusing?  If not, does this
	mean that two clicks are required when preparing to type into a
	text window, one to set the focus, and another to set the insert
	point?  [[ That is one big reason why your moderator really hates
	click-to-type.  Usually one click does both, and that's usually
	not what I want.  --wnl ]]

Well, that's enough for now.  Any other thoughts?

Chuck Musciano
Advanced Technology Department
Harris Corporation
(407) 727-6131
ARPA: chuck@trantor.harris-atd.com

hankg@att.att.com (01/12/89)

In reply to Chuck's posting:

>Unfortunately, I think that Sun View applications will not run under the
>X11/NeWS server without being recompiled as View 2 applications.  I am
>seriously considering creating an X toolkit that is identical to Sun View,
>so I can seamlessly port my tools to X11/NeWS.  Anyone out there have such
>a beast?...

1')  The conversion of SunView to X11/NeWS is a nice idea but as you
implied, it could take awhile to develop.  From what I can see, there is a
current preference toward development in X, because of portability.
"Widget" editors might save time in developing "XView", but I know little
at this time about various X packages available.

>2) Several times during the SUG Open Look seminar, when asked why a
>certain feature was or wasn't included, the answer was based upon
>"psychological reasons" or attempts to "reduce mouse motion".  So,
>I asked Hoeber exactly what cognitive psychological analysis Sun
>had done to measure the interface.  The reply: none. 

2')  This may or may not be totally true.  The group that I work as system
administrator and developer for is "User Interface & Quality Planning".
These guys are PhD's in Behavioral Sciences and Psychology, and have had
some involvement/discussions with Sun over user interface issues.  I have
not seen any actual "Candid Camera" type testing of Open Look going on
here, but there is a proposal in the works for us to do "usability
testing" on Open Look in 1989.

Also....

I don't use a Mac much either, and don't use Presentation Mngr.  I had the
impression that in Open Look Sun is reducing the functionality of their
3-button mouse to agree with a Mac or HP system mouse, and this was not a
good impression.  I guess no one likes having their semi-automatic usage
responses recoded.

Open Look's menus with large buttons spaced rather far apart sort of irks
me.  The stacked button's shading line is not as easy to determine as the
arrow used in current stacked menus.  

Finally...

This concept of setting a User Interface standard "look and feel" perhaps
is going too far.  There are essential parts of cars that look and feel
similar, but everyone enjoys various design changes and flairs.
Personally, I would not want to be the UberMan who says, "This is the best
way to look and feel."   Will Open Look be able to remain truly open, or
will we be locked in and flogged into Open Look?

Hank Grebe
Sun System Administrator, User Interface & Quality Planning 
AT&T Bell Labs
Holmdel, NJ  07733
(201) 949-2549
..!att!hound!hankg

chuck@trantor.harris-atd.com (Chuck Musciano) (01/18/89)

My question: what cognitive analysis was done on Open Look?

Hank Grebe replies:

> I have not seen any actual "Candid Camera" type testing of Open Look going on
> here, but there is a proposal in the works for us to do "usability
> testing" on Open Look in 1989.

Isn't this great?!  Design the interface, release the interface, call the
interface the "standard", and then decide to test it a little bit.  The
more I hear about Open Look, the less I like.

Now that OSF has picked some variant of Presentation Manager, isn't the
future looking bright?  Open Look from ATT/Sun, or PM from everybody else.
Sigh...  That X version of SunView is looking better all the time.

BTW, I was informed by someone at Sun that SunView binaries will run under
the X11/NeWS merge, so all is not lost.

Chuck Musciano
Advanced Technology Department
Harris Corporation
(407) 727-6131
ARPA: chuck@trantor.harris-atd.com

rprobst@sun.com (Richard Probst) (01/21/89)

> Unfortunately, I think that Sun View applications will not run under 
> the X11/NeWS server without being recompiled as View 2 applications.

Let me correct this misimpression before it starts to spread.  The
X11/NeWS server provides full and complete binary compatibility for
SunView applications.  You do not need to recompile.  Your SunView
applications will run unchanged on an X11/NeWS screen.

You may choose to convert a SunView application into a View2 application,
which will require some editing as well as recompilation.  The benefits of
converting to View2 are:

     -- A View2 application can be run on any machine on the network,
	displaying its windows on the workstation you are sitting at;
	this allows you to distribute the workload, and to develop
	networked applications without writing any networking calls.

     -- View2 is an X11 toolkit, so View2 applications will run on
	any X11 server.  On a heterogenous network, with both Suns
	and other workstations, you could run a View2 application on
	a Sun and display its windows on your machine, even if your
	machine was not a Sun (as long as you were running X11).

     -- In the first release, View2 implements a preliminary version
	of OPEN LOOK (basically, it implements the OPEN LOOK spec
	as it went out for industry comment in July; the spec has
	changed since then, based on some of the comments received).
	In the next release, View2 will be certifiably OPEN LOOK.

If these benefits are attractive, you should convert your SunView
applications to View2.  But if you do not need distributed window
operation, or heterogenous window operation, or OPEN LOOK, then you may
decide not to convert.  Either way, X11/NeWS will support you.

	--Richard Probst	(rprobst@sun.com)

ian@dgp.toronto.edu (Ian S. Small) (01/31/89)

> The X11/NeWS server provides full and complete binary compatibility for
> SunView applications....

Gadzooks, this is frightening.  I thought SunView took a lot of space.
The memory taken out by this newfangled windowing system when it starts up
must be truly awesome on a Sun OS 3.* machine.

awm@uunet.uu.net (Aled Morris) (02/11/89)

>> The X11/NeWS server provides full and complete binary compatibility for
>> SunView applications....
>
>Gadzooks, this is frightening.  I thought SunView took a lot of space.
>The memory taken out by this newfangled windowing system when it starts up
>must be truly awesome on a Sun OS 3.* machine.

Not quite.  The way I undestood it was that old SunView binaries will
continue to run, since the SunView 1 library is built on top of Pixrect.
Since the Sun product-version X/NeWS server is also built on top of
Pixrect, the two happly cohabitate on the same screen.  I believe (though
I've never seen it) that the current NeWS product (1.1?) has this property
also (i.e. with NeWS running, you can still run old SunView code).

The effect on the screen is that the SunView window is always the "top"
window in the heirarchy, it kind of floats on top of whatever window
system you are running underneath.  I don't understand Pixrect properly
(and I don't know how the X server Sun ddx code is implemented), but my
guess is that there are rules for "well-behaved" Pixrect applications
which if you obey them, you will automatically get this backwards
compatability.

The MIT X server for Suns does not obey these rules, so you don't have
full compatability with SunView code.  If anyone out there knows how to
fix this, I for one would be extremely grateful!

In the meantime, fire up your MIT X server, and to an xterm, type

	setenv WINDOW_PARENT /dev/win0
	shelltool &

and have fun!

Aled Morris
systems programmer

    mail: awm@doc.ic.ac.uk    |    Department of Computing
    uucp: ..!ukc!icdoc!awm    |    Imperial College
    talk: 01-589-5111x5085    |    180 Queens Gate, London  SW7 2BZ