Macintosh CU-SeeMe Change History README file
11-1-95

by Dick Cogger

This document collects the changes that have occured to CU-SeeMe, 
through the various releases for the Mac, since the start of the 0.70 
series.  Most recent changes are at the top of the file.

-----------------Changes in 0.83b3

1) New version number handling has been implemented, that separates 
protocol version from application version, and allows more flexibility 
in the application version display.  On the info panel under a remote 
window, you will see information on the source and date as well as 
version number.

2) A bug was fixed with outgoing audio bandwidth not being counted 
against the rate cap, which meant that video and other outgoing data 
would not slow down to make room for audio.  It should now be 
automatic for senders to reduce video (if needed to stay under the 
cap) whenever sound starts being sent.

3) A bug was fixed with audio slider controls getting stuck whenever 
moved outside the valid range.

4)  A bug causing crashes on abnormal exits was fixed.

5)  Audio being sent with 7500/8500 models should now be handled 
properly.

CAVEATS

*** Remember, with the 7500/8500,the video digitizer doesn't work (to 
capture to an offscreen buffer as CU-SeeMe needs) if virtual memory is 
on.  We havn't figured out what we're going to be able to do about it, 
beyond buying more memory.  If you turn VM off, 83b3 (and 83b2) works.  
RamDoubler also prevents digitizing.

*** Also, with the 7500/8500 CU-SeeMe and OpenTransport aren't 
compatible with each other.  We are working on making CU-SeeMe 
compatible with Open Transport (and hope that Apple is also working on 
make Open Transport compatible with existing MacTCP-based 
applications).  Meanwhile, do the following: (1) Install OpenTransport 
COMPLETELY, (2) Remove from the Extensions folder the libraries named 
"Open Tpt Internet Library" and "OpenTptInternetLib" and remove the 
TCP/IP control panel, (3) Make sure that you have MacTCP 2.0.6 or 
later (and for dialup, MacPPP 2.1.1SD or later).

*** There is a bug that can cause a crash if you attempt to set a 
nickname to a domain-name rather than a dot-notation IP address.  
(Domain-names are OK to type in to the Connect dialog.)

 
-----------------Changes in 0.83b2 

1) Talk should work better (no more lost messages), and SlideWindow 
should be working again.  Fixed an Auxdata bug that would prevent 
completion of large items, and degrade performance for all items has 
been fixed [Problem was that many aux data packets, including all 
retries, were being rejected by recipient for having incorrect 
header.packetLength]

2) Fixed an audio playout bug that would cause audio to cease after 
5-10 seconds on machines with a heavily burdened cpu is fixed [Problem 
was faulty queue management for audio playout buffers that would fail 
under some combination of routines interrupting each other.  Audio 
playout buffer management has been substantially reorganized].

3) Apple's internal video for the Performa 630 (and some other models 
as well) should now work properly [Problem was that it cannot write 
8-bit gray scale video to memory, only 16-bit color.  Solution was to 
follow 16 bit frame grabbing with an 16 -> 8 bit translation.  BTW, 
this experience confirms the validity of our "supported list" 
approach, as the Apple digitizer would crash if we asked it to grab an 
8-bit gray scale frame...  despite returning no error from the 
VDPreflightDestination call that is supposed to test the validity of a 
given configuration.  Without our "supported list" logic, CU-SeeMe 
would have been unuseable even for receive-only on this class of 
machine].  Probably the 7500 and 8500 PowerMac video is not covered 
by this fix, but we have not received machines of these models yet.

4) A bug that would sometimes cause remote video to be scrambled when 
the source resolution was changed with a connection ongoing has been 
fixed [Problem was the receiver being configured for the wrong size.  
Interestingly, the problem appeared most frequently with a QuickCam, 
though it had no direct relation to the QuickCam, and would appear 
whenever a conference management message with the new configuration 
arrived before a video packet did]


-----------------Changes in 0.83b1

1) some minor interface tweaks.

2) fixed bug causing connection time out followed by immediate resumption,
manifested as participant list shrinking to none, then growing back.

3) fixed bug with wrong names appearing on window titles.

4) fixed bug with audio output level not working on 68K machines. 

5) fixed bug with "please increase your memory to 2100K...." dialog
mysteriously appearing at times in version 83a7.

6) some help with the recently reported problems with people
masquerading as others in the Talk window.  Now, you are not allowed
to change your name while connected.

----------------Changes in 0.83a7

1) The algorithms used for measuring and responding to packet loss have been
improved (they will be fully effective only when used with Reflector 4.00b3, 
available later running at the usual address at Cornell).

2) A new option for the Participant List has senders both hidden and showing
combined in the same part of the list.  They won't move around as you open 
and close them.

3) General cleanup of the UI, adjustment of gray shades, etc. has been done.

4) Active squelch pointer is now highlighted in audio window (upper one active
when audio incoming from the net, lower otherwise).

----------------Changes in 0.83a5

1) Calibrations in the audio Panel. There are now db scales on the level 
meters. Zero db on the meters should indicate maximum modulation of the 
digital envelope (+127, -128 in 8bit).  This level of modulation should 
represent an input signal of 2v RMS or +-2.83v PP input on a PowerMac, if 
you have the input gain set to 0 db also.  These calibrations may not be 
exact (yet), but we're working on it.  The output meter should show the 
content of the received digital samples and will not be affected by the 
setting of the output level control.  When we get everything calibrated 
correctly, an output setting of 0 db (with input also set to 0) should 
give unity gain through the whole system: that is, the level at the line 
output of the receiving machine should equal (slightly delayed) that of 
the input at the sending machine.  We are calibrating for PowerMac's 
since they have line level inputs and outputs.  Earlier Mac's with low 
level (mic) inputs would need an appropriate attenuator.  Signals which 
go above 0 db represent overmodulation or clipping.  Ocaisional clippiing 
of speech is usually acceptable.  We're still futzing with the damping 
for the meters and with access to extra controls, etc.

2) Refinements in Audio Window appearance: When in PTT mode (Push To Talk), 
the squelch thresholds do not appear.  When not in PTT, the labels on the 
button change and the button is "pushed" automatically when you break squelch.

3) To hear just one participant, Option-Click their speaker icon (instead 
of Command-click).  Option-Click again to restore former conditions.

4) Button bars on windows are constructed offscreen, so the annoying flicker 
is gone. 

---------------Fixes in 0.83a4

1)  crash trying to draw audio window icons or drawing wrong audio window
icons is fixed. Also, icons "push in" properly again.
2)  crash upon startup in non PTT mode is fixed
3)  crash upon startup without digitizer is fixed
4)  a non-functioning digitizer is treated the same as no digitizer (e.g.,
instead of error exit on Mac AV with monitor set too deep, the program
continues in receive only mode)
5)  fixed bug that prevented display of outgoing data rate on
digitizer-less machines
6)  point-to-point connections no longer require that at least one party
have video

Regarding 4, there is still a problem with PowerMac AV's which causes them
to crash, rather than returning an error code that CU-SeeMe can respond to,
when trying to access the digitizer when the monitor is set too deep.  This 
problem may be fixed with an upgrade to the System software, but Tim's 
7100/66AV with System 7.1.2 does crash.


----------Version 0.83a3 has the following updates:

* Save Window Positions now works properly again.

* The button bar icons, background color, recent/last speaker indicators 
    are all in sync with those in the Participants List.

Outstanding items/issues at this time (besides various bugs) are:

* We're still fiddling with the appearance of icons, etc.

* The button bars are updated unnecessarily (you can see them flicker)

* VU scales in the Audio window are updated unnecessarily (if you put the cursor 
   on the edge of the scale, you will see it flicker)

* The audio input gain control does not operate smoothly.


-------------Version 0.83a2 included:

* Clear/Enter key works again properly for burnt text-in-video
* Participant List no longer condenses long participant names
* Participant List doesn't get lost in invisible desktop space anymore
     (checks for invalid position)
* CU-SeeMeª Talk now works... (it was AutoDoubled accidentally by the server)
* Color schemes matched
* Lots of other small optimizations and features

* You can expand the Audio Window by clicking in the Grow Box (upper right 
      corner) to access more options...
    
--------------Version 0.83a1 original notes:

        Major changes are the following:

1. There is a new algorithm and format for loss reports and more
communitcation with the new version of the reflector.  You now have
parameters to set to limit your overall receive bw.

2. There is a participant list (Show Participant List from the Participants
Menu) which will show information about senders and lurkers and provide
control of audio, access to statistics, IP address, etc.  There are several
options in the Prefs dialog to control how the list is resized, and you can
twiddle small triangles to show or hide portions of the list, in a fashion
similar to finder list views.  The Participants list essentially consists
of the button bars for various users, giving you the audio controls and the
"eye-con" that shows who is watching your video stream.  Enhanced
indication of who is sending audio, who was the last speaker, who has
recently been speaking, etc. is provided as well.  You can tell the
difference between someone having their audio completely off versus just
off for you (big X, little x).  Note that details of these UI changes are
still being adjusted; there will probably be additional alpha releases over
the next week or so.

3. The Disconnect item in the Conference Menu now shows the IP of the other
end of your connection, either user or reflector as the case may be.

4. There is a new Audio window with level indicators for both your mic (as
before) and for incoming audio.  Also there is an input gain control to
adjust the sensitivity of the mic and an output volume control to adjust
the speaker level.  The input level meter indicates the level to be
transmitted and is affected by the setting of the input gain as well as by
how loud you speak.  The output level control shows the level received in
the packets (it should show the same level the person sending saw on their
meter) and is unaffected by the output volume control--that just controls
how loud it sounds at your end. Other features are experimental: a dual
squelch control allows you to raise the squelch level when sound is being
received from the net.

5.  You can use the Control key as a push-to-talk activator instead of
having to mouse to the right place in the audio window.  CU-SeeMe needs to
be in front, however.

6. If you hold down (instead of clicking) on someone's mic icon, it
functions as a private-to-them push-to-talk.  If you Command-click someones
speaker icon, instead of turning them off, it turns everyone else off.

7. Lurker windows can be opened, and you get a local window if you are a
lurker (just title bar and button bar) to give access to various controls.
In particular, a lurker can access the control panels under the local
window for audio and receive bandwidth.

8. The Participant List will show up to 24 participants (up from previous
limit of 16 on the menu), so you can keep track of more senders/lurkers on
a busy reflector.  Be warned, though, that the conference management
messages for 24 users can pretty well use up a 14.4 modem link.

9.  QuickCam Audio works (yea!), although the sound is not the highest
quality.  QuickCam provides 5khz sampling, and the present implementation
simply interpolates that up to 8khz so it will go thru the encoders as they
stand.  Puffing up the bandwidth utilization at the sending end by 8/5ths
is clearly a sub-optimal idea, but to interoperate with everything it was
seen as a necessary mode and a starting point.

Larry Chace, Tim Dorcey, Jeff Han, and John Lynn contributed to the code
and design for this release.


------- 0.80b1 ---------

Slide Window stuff:

1.  Slide Window AuxData type has been changed from 2 to 'CUsp' 
(transparent to user except different versions do not recognize each 
other-- i.e., 0.80a3 is obsolete.)

2.  When you send a slide, the AuxData routines keep it available for 
receivers to request retrys of missing pieces until one of the 
following conditions is true:

  1.  A new slide is sent
  2.  MAX(30 seconds,twice original transmission time) elapses
  3.  The slide window is closed
  
Previously, only 1. was controling.

3.  Slides will not be received from a participant unless you are also 
receiving video from that participant.

4.  The slide window menu item now alternates between "Open slide 
window" and "Close slide window"

5.  The vertical dimension of the slide window is adjusted to fit on 
the monitor.  Transmitted slides are the same size as is displayed, 
and will be positioned at the top of larger destination slide windows.  
Slides arriving from a larger source are truncated at the bottom.

6.  The slide window large cursor will not appear and no remote 
pointing/queuing will be allowed unless the item has been sent at 
least once. Previously, it appeared to work, but nothing was sent.

7.  When you leave Local mode, the slide that would have been queued 
had you not been in Local mode is queued.  In other words, you catch 
up with the current state of the conference.

8.  Fixed an anomaly in the button behavior that would occur if a new 
slide arrived while you were in Local mode, but yet had no slides in 
your stack

9. If you try to open SlideWindow without Quicktime installed, it 
will not be allowed, since SlideWindow uses the QT JPEG compressor.  
Even if you are receive-only, you need QuickTime in order to receive 
slides.

Miscellaneous:

1.  The application beeps if a connection request is received while 
you are in Waiting state (regardless of whether the acceptance dialog 
appears)

2.  The application clicks whenever a new participant joins the 
conference, subject to a (new) preference item.

3.  Video windows do not automatically come to foreground when they 
open, solving problem with Talk window typing being misrouted

4.  Menu item for audio window brings it to the foreground (instead of 
just making it visible).

5.  Reflector messages terminate the connection immediately (instead of
waiting for user OK).

6.  The reception hold down interval (during which time packets from a 
recently closed connection are ignored) has been increased from 30 
sec.  to 60 sec.  Some comment on this is necessary to insure that it 
doesn't confuse people in point-to-point testing situations.

7.  Audio squelch level is saved in preferences

8.  Frame-Differencing and square decompression routines in PPC 
version have been modestly optimized.

9.  Packet sequence number wrap-around is now allowed (to account for 
nv's 2-byte seqnum).

10.  The kbps display in the rates bar now includes all packets sent 
or received on the CU-SeeMe port, with the exception of packets that 
are received from an IP address for which there is no rate bar to 
display it.  Previously, only video was counted in the outgoing 
direction, and only _valid_ video, aux data, and OC packets were 
counted incoming.  E.g., out of sequence packets, video packets for a 
closed window, etc.  were not previously counted, nor was audio.

11.  The transmission rate cap applies only to video+aux data, so OC 
packets and audio packets may cause the outgoing kbps to consistently 
exceed the cap.

12.  The pull down stats display and the packet loss statistics that 
are used to adjust the rate cap are based on video + oc packets, and 
are only accumulated when video is being received.  This approach is 
required since aux data and audio data have different sequence number 
spaces.  This is essentially the same as before, except that there was 
a logic error that would sometimes allow a packet to enter the byte 
count without entering the packet count.  Hence, the average packet 
size may have been over-estimated before.

13.  Copyright and Permission notices have been modified: Contained in 
the application's About and also in the README.0.80.basic...  file, it 
has been changed slightly: permission to modify has been deleted (not 
very meaningful with binary only, will be spelled out for Plug-In 
example sources), permissions granted explicitly only for the current 
version, reserving the right to modify permissions with future 
releases.

------- 0.80a3 ---------

1.  A fix to the pop-up in the local window control panel for 
selecting digitizers which appeared in 80a2 (and in 70b13).  The 
previous version *may* have introduced a freezing problem (although it 
appeared to work fine), especially when using System 7.5, especially 
on PowerMac's (we think).  This fix is also in 0.70b15.

2. Please do not connect to "Self" as it will probably hang up your 
machine.

------- 0.80a2 ---------

Note that the Plug-Ins are new and won't work with 80a1 and the old 
plugins won't work with 80a2.

What's Different:  

1) QuickCam support--The ability to select vdig (so you can use 
QuickCam) is brought forward from 0.70b14.  However, the QuickCam 
won't produce a 640x480 image, so the SlideWindow is less useful than 
with a normal video source.

2) Delta-Mod 16K audio -- Charlie Kline's own 16Kbps audio encoding 
which first showed up in 70b13 has been brought forward.  Remember 
that vat users will not be able to decode this signal; just Maven and 
CU-SeeMe for now.

3) The four required Apple Events are now supported, thanks to Aaron 
Giles.

4) DNS names are now supported in the connect dialog, but not yet in 
the Nickname file.  A further release will have support for url format 
nicknames and saving bookmark files containing a url which will launch 
CU-SeeMe and connect to the site pointed to by the url.  Apple events 
will also work for launching/connecting, so it should be possible to 
set up web brousers to establish CU-SeeMe sessions.  But we havn't had 
time to wring this stuff out for this release.

5) There is a new preference item to open windows automatically when 
new video sources are discovered, which was previously the standard 
behavior.  The new default is NOT automatic.  That means that you 
would need to check this box to get the old way of working.  The 
purpose of this change is to prevent naive first-time users from 
hooking to a busy reflector via a modem link, with default settings, 
and getting a flood of 8 windows most of which can do nothing but jam 
up their slip server.  For the rest of us, it will allow you to get 
the folks on the screen you want and then not be annoyed by additional 
windows ceaslessly popping up.  The full utility of this mode must 
await implemntation of the "Participants List" window which will have 
an entry for each particpant who is not displaying in a window.  This 
option only applies to reflector connections.  In 2-party, available 
video is always opened.

6) There is a new preference item to disable direct screen drawing, 
and omit black borders if direct screen drawing is disabled.  Also, 
black borders are 2 pixels wide rather than 4.  For those doing demo's 
with powerbooks and projectors, wanting to run in Mirror mode, this 
choice should allow the desired operation.

7) Default "Change Tolerance" is now 30 instead of 24.  The result 
should be just a little less bandwidth, but little visible impact.  On 
AV Mac's, I find that a value of 34 seems tolerable.

8) Change in IP address no longer wipes out old Preferences.

9) Preference file handling now allows compatibility across versions, 
so that we can add additional fields without destroying old 
information.  Running an old version will, however, wipe out any new 
preference fields that it is not aware of.  This will require a minor 
update to the 0.70 series (look for 0.70b15) to allow seamless 
switching between .80 and .70 series (except that running.70 will 
erase new preference items 5) and 6) above)

10) The "Local/Remote" button on slide projector behaves differently: 
it always displays as Local, but when you click it highlights (rev 
video) meaning you are *in* local mode.  When it is normal black on 
white, it means you could click it to *go into* local mode.

11) The Aux Data facility has been improved in several ways.  It can 
now transmit various items in an interleaved fashion, so that a large 
item will not delay the transfer of other items.  It will repeat very 
short items a few times to enhance the probability of correct 
reception.  It will avoid sending unnecessary control information.  It 
will more efficiency handle error recover for multiple recipients.  
The plug-in interface now keeps Aux Data items around for 15 seconds 
after they have been sent greatly enhancing the probability of the 
items being received correctly.  In our testing, the result is that 
the Talk window responds quickly and reliably, even when one or more 
slides are being transmitted.

------- 0.80a1 -------  OVERVIEW: new in 0.80  

1) A "SlideWindow" facility with which you can capture 640x480 stills,
transmit to all participients (who have SlideWindow open) and then point
with the mouse and have everyone see the pointer.  You can also switch
among a number of slides and have the remote partipants see the same thing.
SlideWindow is an "AuxData Application" which is invoked with the SlideWindow 
menu item on the Conference menu.

2) An "Auxilliary Data Transport" providing reliable one-to-many delivery
of data constituting auxilliary materials for sharing in a conference.
Data is transmitted as "Items," with an item being defined as a string of
bytes contained in memory. The AuxData transport takes care of packetizing
and any neccessary retrys to ensure reliable delivery.  This transport is
used by the built-in SlideWindow application or by plug-in AuxData
applications (see below).

3) A Plug-In interface so that seperate modules can add functionality 
to CU-SeeMe.  There will be (soon but not with the alpha release) a 
"Function Module Developer's Kit" consisting of documentation and 
sample plugins and a CodeWarrior project with libraries to facilitate 
development.  If one or more plug-ins is in the same folder 
as CU-SeeMe at launchtime, a menu will be added (for each).

4) A "talk" plug-in (in source as an example) for sharing
typed messages as with IRC.  The Talk module implements a "chat" window 
that allows exchange of text messages. 

5) An AuxData tracing tool for examining data flows when debugging or
testing AuxData applications. The AD-Trace module is a testing or debugging 
tool that shows all "AuxData" traffic.

6) Also look for a new version of the reflector which will prevent
sending of AuxData to participants who are not running the appropriate
AuxData application (or who don't have it).  This version is currently
running at Cornell at the usual address, 132.236.91.204.  You can use this
reflector or you can make point-point connections to test the AuxData
applications provided with the alpha 0.80. If any of the alpha-test group
can run the new reflector, currently 2.50b1, please do and advise this
group. With luck, there will also be a further enhanced reflector 
out soon, with additional facilities for interoperating with vat and 
nv. That version will also support AuxData transmission.

------- 0.70 b15 changes ---------

1.  A fix to the pop-up in the local window control panel for 
selecting digitizers which appeared in b13.  The b13 version *may* 
have introduced a freezing problem (although it appeared to work 
fine), especially when using System 7.5, especially on PowerMac's 
(we think).

2.  A fix to a problem with handling Preferences files, especially 
when changeing back and forth from 0.70 to future (alpha) versions.  
This fix erroneously announced as being in b14, previously.

-------- 0.70 b14 changes ---------

1.  There is an additional audio encoding-- Charley Kline's delta-mod 
16K bps encoding (thankyou again, Charley).  This should work much 
better with low bandwidth links, such as 64K ISDN links, and possibly 
with 28.8K modem connections.  However, compared to the DVI 32K 
encoding, assuming no packets lost in either case, the sound quality 
is noticibly poorer, although speech is entirely intelligible.  You 
will not be able to use this encoding to converse with anyone using an 
earlier version of CU-SeeMe.  You will not be able to interoperate 
with VAT.  You should be able to use it with recent versions of Maven.

------- 0.70 b13 changes ---------

1.  Support for the Connectix QuickCam (Yea!), the $100 combined 
camera and digitizer that plugs into your serial port, including an 
addition to the control panel you see under your own window that 
allows you to choose which of available digitizers to use.

2. Correct strange behavior when using a conference ID greater than 
255.

3. Correct behavior when your own video is "paused" -- it was 
sending at much too high a rate.

4. Internal change dealing with packet sequence numbers and 
interoperation with VAT for periods longer than a couple hours.

NOTE:  This version will replace existing "CU-SeeMe Preferences" 
file, causing all settings to revert to default values.  Before 
running this version, you may wish to take note of your current 
settings (e.g., Compression Tolerance, your name if not easily 
recalled, etc.).


------- 0.70 b12 changes ---------

1. Fix a bug introduced in 0.70b10 that caused audio connectivity
to be flaky.  In other words, some participants would receive
your audio while others would not.

0.70 b11 changes:

1. Fix a bug in the processing of audio which potentially could be the
cause of random crashes.

2. Allow incoming audio packets to continue to be processed even while
the mouse is held down in the menus.  For example, this allows incoming
audio to continue to play out while you are looking through the list of
participants in the Participants menu.  Audio will continue to play out for 10
seconds at which point your connectiong will time-out in the reflector and
it will stop sending you CUSM packets.

------- 0.70 b10 changes -------

1. Nicknames >30 char no longer cause a crash on startup (but will be
truncated to 30 characters).  With the previous b9, the workaround for
this bug was to avoid long nicknames.

2. Old Prefs files no longer cause crashes when opening a sound channel
(as when clicking Push To Talk).  With the previous b9, the workaround
was to through away the preferences file and be sure to quit cleanly
after using audio to save a good prefs file.

3. There is a small change to the OpenContinue message (CU-SeeMe's
conference control protocol messages) as to the indication of
changes-- result is much less work for the reflector as people come
and go on a busy reflector.

As before with bug fixes, there should be no user visible changes
(apart from lack of crashes for the unwary).  Reflector operators
may notice many less log messages when folks have converted to this version.

------- 0.70b9 changes --------

1. One case of "push to crash" related to wrong initialization of a
sound channel.  Not clear whether this fixes all cases.

2. A bughalt on a Mac IIcx because it was unable to initialize 4 sound
channels.  CU-SeeMe will now be happy if it can initialize at least one.
It still trys to open 4 sound channels at initialization, but if it gets
an error return after the first, if will settle for what it has.

3. Bughalts will now not send you to the dumper (or MaxBug) but will give
a dialog with a message and then quit cleanly when you click OK.

4. There was a problem where the MinKbps control under Transmission
under the Local window worked improperly.  In particular, if the MinKbps
were set to a value greater than 100, the first digit might be invisible.
It could then appear that Min was, say, 10 but the Max could not be set
below 110 or 210.

-------  0.70b5  ---------

Be sure you use SoundManager 3.0 extension with Audio capable
CU-SeeMe (but NOT on the PowerMac).

        CU-SeeMe0.70b5 is now available on gated.cornell.edu /pub/video
for anonymous ftp.  ***WARNING***  This is beta software.  It is
based on the previous 0.70b1 and contains no new features, only bug
fixes.  Several bugs in the area of audio have been fixed, and we do
not expect that we have introduced new problems, but software folks
always say that and aren't always correct, so beware.  There is at
least one crash producer still in this code.


------ New Stuff in 0.70 -------  0.70b1

The principle enhancement in 0.70 versions is the inclusion of
Audio, based on Charlie Kline's Maven code.  It works with the new
reflector and supports multiparty operation.  BUT.  ***Now Hear This***
(and tell your friends): AUDIO WILL NOT WORK CONNECTING WITH A MODEM!!!
I know it would be great if it did, but it just doesn't.  You need at
least 32K bits/sec just for the audio (when speaking or listening) and
modems only do 14.4K.

If you decide to be a tester, watch out for the following:

* Don't run CU-SeeMe0.70b1 and FTPd at the same time.  At least on a
Quadra 840AV it crashes after a while.  Don't know if b5 changes this.

* Problems have been reported with LC's and Quadra 650's using Spigots.

* One particular Powerbook 180 and one PowerMac 8100 had crashes, but
others worked fine.

* Results trying to use MacRecorder (on older Mac's, like the venerable cx
and ci, which don't have built-in audio) have been problematic.  We've
ordered a MacRecorder and will try to come up with a clear understanding.
Any info/experience would be welcome.

* Use the new reflector.  If you hook to a reflector, and it's running the
new code, you will probably get a Message Of The Day, telling you whatever
the reflector operator wants you to know.  70bN will probably work with
the old reflector code for video, but there may be problems with the audio.

Shaking out these model-specific problems may take a while, since it's
really hard for us to reproduce all the variations here.  If you do
experience crashes (and still want to continue), please try without any
extra inits, or perhaps without other apps running; we have had problems
with the moire screensaver, for example.



REFLECTOR        There is also a new version of the reflector, which you
need to run to have the audio properly handled.  (Note: we are one release
further to b2 compared with the version released last week.  You need this
new version to support the new MOTD feature in the reflector/client.)  It
will also permit the use of stand-alone Maven.  It is on the usual
anonymous ftp site, gated.cornell.edu /pub/video.

AUDIO        Audio is controlled from a small window and works just like
Maven.  You can choose Push-To-Talk mode (PTT) or set a squelch level with
a small pointer to the left of the audio-level bar.  There are boxes in
the audio window to turn receive and send on and off entirely.  Or you can
turn off audio from a particular particpant by pushing the speaker button
under their video window (they will get the mic on your window X'd to
indicate you won't hear when they speak).  Also, you can choose to turn
off receiving from standalone Maven clients who connect to the reflector
(a temporary hack until we have a list-window for participants not
displaying video).
        If you push the mic button on someone's window, you establish a
private talk channel to them.  Push again to go back to general conference
audio sending.

CONFERENCE MENU    The Conections menu has been renamed to "Conference"
since it now includes an item to open the Audio Window (if you've closed
it).  In future, there will be more conference-management related items on
this menu.  Of course, as time goes by, we may need to re-organize the
menus to make sense of future developments which aren't even gleams in the
eye today.

THE EYE    Also on remote window button bars is an eye-con (sorry,
couldn't resist).  If open, it means that the person in the window is
looking at you-- or actually that they have a window open on their desktop
showing your video stream.  If they close the window, the eye at your end
closes.  When the eye is open, it looks a little sinistre, some folks
think.  Icons are still being edited.

MOTD    With the new reflector, you can set up a message which will be
displayed to folks as they connect.  (If they use 0.70 or later.)  If you
operate a reflector, you might want to use this facility to explain
policies, etc. that apply to use of your reflector.

SAVE WINDOW POSITIONS   New on the File menu is an item you can use to
save current window positions.  Until you use it, they open tiling left to
right.  You probably want to move them around while you remember the order
they came up and then save.  The Participants menu lists the order that
windows open as you initially connect to a reflector.  You can go back to
defaults by chucking your preferences file.  This facility will be
improved to make it easier to get the results you want.

THE BLACK BORDER  If part of the video portion of a CU-SeeMe window is
covered, the program uses QuickDraw to move video into the window.
Otherwise, optimized routines draw directly to the screen.  Particularly
on a slower Mac, drawing with QuickDraw is very slow.  So you will know
when this is happening, a black border shows on such a window.

PREFERENCES LIMIT ON NUMBER OF WINDOWS You can set a preference for the
maximum number of windows you want to show at one time (default 8, set
between 2 and 8), including your local window.  This feature will help if
you have limited screen real-estate.

CONFERENCE ID    New on the Connection dialog is a box for conference ID,
default of zero.  The new reflector will reject all participants with the
wrong number, and it can be configured with a conference ID other than
zero.  If both reflector and participants go with the defaults (zero), it
works the way it always has.

-------------Basic CU-SeeMe Info---------------

Get the latest file labeled CU-SeeMe.README.txt.<date> for a 
consolidated collection of information on CU-SeeMe.

CU-SeeMe was initially written for the Macintosh by Tim Dorcey with design
assistance and sponsorship by Richard Cogger of the Advanced Technology
group in the Network Resources division of Cornell University's Information
Technology department (CIT).  Important early contributions came from:
Cornell University Medical Colleges (CUMC), Scott Brim, and John Lynn.

Since Oct.  1, 1993, the CU-SeeMe Project receives funding from the
National Science Foundation.  A very significant collaborative effort at
Cornell University Medical Colleges (CUMC) is contributing substantial
expertise and code.

This material is partially based on work sponsored by the National
Science Foundation under Cooperative Agreement No. NCR-9318337.  The
Government has certain rights in this material.

*****The fine print*****

Copyright 1993, 1994, Cornell University

Cornell hereby grants permission to use, copy, modify, and distribute this
program for any purpose and without fee, provided that these copyright and
permission notices, and those of Cornell's collaborators included below,
appear on all copies and supporting documentation, the name of Cornell not
be used in advertising or publicity pertaining to distribution of the
program without specific prior permission, notice be given in supporting
documentation that copying and distribution is by permission of Cornell.
CORNELL MAKES NO REPRESENTATIONS OR WARRANTEES, EXPRESS OR IMPLIED.  By way
of example, but not limitation, CORNELL MAKES NO REPRESENTATIONS OR
WARRANTIES OF MERCHANTABILITY OR FITNESS FOR ANY PARTICULAR PURPOSE OR THAT
THE USE OF THIS SOFTWARE OR DOCUMENTATION WILL NOT INFRINGE ANY PATENTS,
COPYRIGHTS, TRADEMARKS, OR OTHER RIGHTS.  Cornell shall not be held liable
for any liability with respect to any claim by the user or any other party
arising from use of the program.

The audio portion of CU-SeeMe was provided by Charley Kline's Maven.

 Charley Kline, cvk@uiuc.edu
 University of Illinois Computing and Communication Services
 28 Feb 1994

 Copyright (c) 1994 by Charley Kline and the University of Illinois Board of Trustees.

  Redistribution and use in source and binary forms, with or without
  modification, are permitted provided that the following conditions
  are met:

  1. Redistributions of source code must retain the above copyright
     notice, this list of conditions and the following disclaimer.
  2. Redistributions in binary form must reproduce the above copyright
     notice, this list of conditions and the following disclaimer in the
     documentation and/or other materials provided with the distribution.
  3. All advertising materials mentioning features or use of this software
     must display the following acknowledgement:
       This product includes software developed by the University of
       Illinois, Urbana and its contributors.
  4. Neither the name of the University nor the names of its contributors
     may be used to endorse or promote products derived from this software
     without specific prior written permission.

  THIS SOFTWARE IS PROVIDED BY THE TRUSTEES AND CONTRIBUTORS ``AS IS'' AND
  ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  ARE DISCLAIMED.  IN NO EVENT SHALL THE TRUSTEES OR CONTRIBUTORS BE LIABLE
  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  SUCH DAMAGE.