[Webteam] Squeak History

Dan Ingalls Dan at SqueakLand.org
Mon May 21 16:36:27 UTC 2007


>I remember being told that the ST-76 image was cloned from the ST-72 
>image by transferring the ST-72 objects (using VMWriter?). So my 
>current squeak image is a direct descendant of ST-72 because there 
>has been no start from scratch with an empty image.
>
>Not one of the cells I was born with 77 years ago exist today. Still 
>-- I'm still me. By the same token, I believe Squeak as a living 
>thing was born as the ST-72 image and is still going strong.
>
>Cheers
>--Trygve

Hi, Trygve -

Your recollection (or your source) is close, but not quite accurate.

We definitely used St-72 to build St-76, but it did not *become* 
St-76 the way St-76 became St-80 and St-80 became Squeak.  St-76 was 
a complete departure from the past, and it was all written down as a 
set of about 35 classes and 500 methods totalling about 20k of code. 
I wrote a compiler in St-72 that compiled that bootstrap into an 
object memory, and that was then tested with the new bytecode VM in a 
debug cycle until it worked.  It was not until a month or two later 
that Ted got the first VMemWriter (now referred to as SystemTracer) 
to work, and this has allowed us ever since to carry objects forward 
into the future even across changes to the object memory.

But don't let me spoil your story -- there are, in fact, a few 
objects that survived even this step from St-72 to St-76, and these 
are the cursors (at least their foreground bits).  Many have come and 
gone since then, but I personally assembled the bits for the 
eyeglasses cursor as octal numbers back in St-72.  A section in the 
St-76 bootstrap process copied the cursor bits out of the St-72 
bootstrap reader itself and into the new St-76 image, so these 
objects survived in the VMemWriter sense from ST-72.

Fun stuff

	- Dan
-----------------------------------------
>On 19.05.2007 19:36, Alan Kay wrote:
>
>>Dan and I have joked that there is probably at least one line of 
>>code I wrote for class Paragraph still there. (I did the original 
>>version in ST-72 -- a one -pager -- but it has been added to and 
>>etc., many times since by others.)
>>
>>However, there is quite a bit of code still in Squeak from the 
>>Smalltalk-80 release version that went to various companies, 
>>including Apple before the Blue Book was written. And at least some 
>>of that was in ST-76, the original version of which was pretty much 
>>designed and written by Dan (the other three contributors for that 
>>implementation were Ted Kaehler, Diana Merry, and Dave Robson). 
>>Larry Tesler soon afterwards added the first of the code browsers, 
>>etc.
>>
>>Cheers,
>>
>>Alan
>>
>>At 10:09 AM 5/19/2007, Brad Fuller wrote:
>>
>>>subbukk wrote:
>>>
>>>>
>>>>On Friday 18 May 2007 11:02 pm, Craig Latta wrote:
>>>>
>>>>      
>>>>
>>>>>>
>>>>>>not forgetting of course that Squeak was announced in september(?)
>>>>>>96
>>>>>>      
>>>>>>          
>>>>>>
>>>>>
>>>>>      midnight pacific time, 1 October 1996.
>>>>>    
>>>>>        
>>>>>
>>>>
>>>>Is the birth date of Squeak Virtual machine or the Squeak Image?
>>>>Though Squeak
>>>>was released in '96, some of the objects in its image must have been
>>>>created
>>>>long before that. Is there a way to find out the oldest extant object in
>>>>an
>>>>image?
>>>>
>>>>      
>>>>
>>>According to Dan's "Back to the Future"  article, the team started 
>>>with an Apple Smalltalk-80 implementation which contained the 
>>>image and VM. Maybe there is still some code from ST-80.
>>>
>>>brad
>>>
>>
>>
>>
>>
>>
>>
>
>
>--
>
>Trygve Reenskaug      mailto: <mailto:trygver at ifi.uio.no>trygver at ifi.uio.no
>Morgedalsvn. 5A       <http://folk.uio.no/trygver>http://folk.uio.no/trygver
>N-0378 Oslo           Tel: (+47) 22 49 57 27
>Norway
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20070521/ae7bd49c/attachment.htm


More information about the Squeak-dev mailing list