Squeak-dev July 1998

squeak-dev@lists.squeakfoundation.org

80 participants
111 discussions

TextConstants in 2.1
by Georg Gollmann 02 Jul '98

02 Jul '98

Another minor observation on 2.1: TextConstants includes CrLf and CrLfCrLf, which is nice :-) Text>initialize however does not set these up - maybe they just crept in via my miscChanges.st. So what is the official content of TextConstants ? Whatever shipped in the distribution image or the result of doing Text>initialize on an empty dictionary ? Georg ---- Dipl.Ing. Georg Gollmann TU-Wien, EDV-Zentrum phon:(+43-1) 58801 - 5848 fax: (+43-1) 587 42 11 mail:gollmann@edvz.tuwien.ac.at http://macos.tuwien.ac.at/Gollmann.html

2 1

Re: Mail Client in Squeak
by James O'Connor 02 Jul '98

02 Jul '98

Ted, Thanks a lot for pointing that out. It will be very helpful. Take care, Jay >Eliot wrote: >>and grab STAMP from the parc goodies in the VisualWorks Non-Commercial >>distribution. It contains a base64 decoder and I have a uuencode >>decoder for it also. No encoders yet. STAMP also manages multiple >>folders etc. I use it as my mail client at work. > >Folks, > Squeak already contains a class Base64MimeConverter in category Collections-Streams. Here is its class comment: > >This class encodes and decodes data in Base64 format. This is MIME encoding. We translate a whole stream at once, taking a Stream as input and giving one as output. Returns a whole stream for the caller to use. > 0 A 17 R 34 i 51 z > 1 B 18 S 35 j 52 0 > 2 C 19 T 36 k 53 1 > 3 D 20 U 37 l 54 2 > 4 E 21 V 38 m 55 3 > 5 F 22 W 39 n 56 4 > 6 G 23 X 40 o 57 5 > 7 H 24 Y 41 p 58 6 > 8 I 25 Z 42 q 59 7 > 9 J 26 a 43 r 60 8 > 10 K 27 b 44 s 61 9 > 11 L 28 c 45 t 62 + > 12 M 29 d 46 u 63 / > 13 N 30 e 47 v > 14 O 31 f 48 w (pad) = > 15 P 32 g 49 x > 16 Q 33 h 50 y >Outbound: bytes are broken into 6 bit chunks, and the 0-63 value is converted to a character. 3 data bytes go into 4 characters. >Inbound: Characters are translated in to 0-63 values and shifted into 8 bit bytes. > >(See: N. Borenstein, Bellcore, N. Freed, Innosoft, Network Working Group, Request for Comments: RFC 1521, September 1993, MIME (Multipurpose Internet Mail Extensions) Part One: Mechanisms for Specifying and Describing the Format of Internet Message Bodies. Sec 6.2) > >By Ted Kaehler, based on Tim Olson's Base64Filter. > >--Ted. > > >Ted Kaehler, Walt Disney Imagineering, R&D >(home) 3415 Cork Oak Way, Palo Alto, CA 94303. voice (650) 424-1070 >http://www.webPage.com/~kaehler2/ >Give me your old Macintosh! I will put it in a first grade classroom. > > > > Jay O'Connor joconnor(a)roadrunner.com "God Himself plays the bass strings first, when He tunes the soul"

1 0

Re: Squeak 2.1 now available
by Tim Rowledge 02 Jul '98

02 Jul '98

OK, all you Acorn users, it should be visible, assuming I got the permissions and so on right, on the usual Sumeru.stanford.edu/tim site. tim PS ARMLinux version coming soon, just as soon as I can get the makefile etc changes to Ian. Oh, and an Itsy version :-). -- Useful random insult:- Mouth is in gear, brain is in neutral. Tim Rowledge: rowledge(a)interval.com (w) +1 (650) 856-7230 (w) tim(a)sumeru.stanford.edu (h) <http://sumeru.stanford.edu/tim>

1 0

PositionableStream>>nextWord
by Martin Pammer 02 Jul '98

02 Jul '98

I would suggest to rename two methods of PositionableStream: nextWord "Answer the next two bytes from the receiver as an Integer." nextWordPut: aWord "Append to the receiver an Integer as the next two bytes." to next2Bytes or nextTwoBytes and next2BytesPut: or nextTwoBytesPut: The reason for this suggestion is that I would like to see a method nextWord to return the next word from aStream in a literal sense. I remember that at least VSE has a method like this !PositionableStream methodsFor: 'accessing' stamp: 'pm 6/14/1998 22:24'! nextWord "Answer a String containing the next word in the receiver stream. A word starts with a letter, followed by a sequence of letters and digits." | first | [self atEnd ifTrue: [^ nil]. self next isAlphaNumeric] whileFalse: []. "skip separators" first _ self position. [self atEnd ifTrue: [^ collection copyFrom: first to: self position]. self next isAlphaNumeric] whileTrue: []. position _ position - 1. ^ collection copyFrom: first to: self position! ! With best regards Martin Pammer mpammer(a)vrz.net VRZ Informatik

1 0

Numeric Anecdote (Nerd Factor 9)
by Dan Ingalls 02 Jul '98

02 Jul '98

Ted Kaehler reported to me that evaluating 2.0e78 asInteger caused Squeak to run out of memory. This has been so for a long time. I traced it down to the failure code for <Float> truncate. <Float> truncate is a primitive that succeeds whenever the result can be expressed as a SmallInteger. If it fails, it returns ^ (self quo: 16383.0) * 16383 + (self rem: 16383.0) truncated Clearly this code goes back to the days of 16-bit pointers. But, hey, it works. Mostly. For fun, I timed the execution for values up to 2^150. As follows... (1 to: 150) collect: [:i | Time millisecondsToRun: [(2.0 raisedTo: i) truncated]] (0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 1 0 1 1 1 1 1 1 1 1 1 1 1 1 10 1 2 2 3 2 2 3 2 2 2 3 2 3 2 3 5 5 5 4 5 5 5 6 15 5 6 5 6 6 10 10 10 11 17 11 10 11 11 12 11 19 12 13 21 27 22 23 22 29 22 21 30 23 25 49 28 27 49 44 54 56 51 56 51 55 79 49 57 67 54 60 99 113 97 94 96 101 111 109 106 102 137 105 127 109 199 214 187 215 197 215 215 219 225 ) I thought about how to fix this. Updating 16383 to Squeak's current SmallInteger limit helped -- it allowed 2.0e78 asInteger to complete, but it was still very slow. It has to call itself N levels deep to deal with a number N times longer than the divisor used. Moreover the code ultimately loses precision, and gets in a deadly embrace between quo: and rem: owing to the limited precision of Floats. This is why 2.0e78 asInteger did not compute. Clearly when the number is very large, we should just extract the mantissa, and shift it according to the exponent. I had already fastened my seat belt to do this when, browsing around in Float, I (re)discovered asTrueFraction by David N. Smith and Luciano Esteban Notarfrancesco. At first it looked as though it had a lot of similar code in it. But after reading it more carefully, I realized that it actually computes exactly the result we need! It's not often that you just happen across a method lying around, referred to nowhere, that will fix a bug you have, and run much faster to boot. So here's the new code: self abs < 2.0e16 ifTrue: ["Fastest way when there may be a fractional part" ^ (self quo: 1073741823.0) * 1073741823 + (self rem: 1073741823.0) truncated] ifFalse: [^ self asTrueFraction. "Extract all bits of the mantissa and shift if necess"] It works! Bit-perfect every time! And never more than a millisecond. Update forthcoming.

1 0

RE: Leads for database application development
by Jeffrey Odell 02 Jul '98

02 Jul '98

I spoke to Kyle this week at Object Expo, Europe. KSC was bought out by another company (sorry, I don't remember who) after having severe financial problems. Kyle went to this company, but shortly thereafter moved to IBM. I have read Kyle's paper and heard his talk on Crossing Chasms. They are both excellent. I'll ask him where the link is, and when I find it, I'll post it to this list. jlo > -----Original Message----- > From: Jarvis, Robert P. [mailto:Jarvisb@timken.com] > Sent: Wednesday, July 01, 1998 8:11 PM > To: 'squeak(a)cs.uiuc.edu' > Subject: RE: Leads for database application development > > > I just remembered a pattern language I came across a while ago. It's > called "Crossing Chasms" and was developed by Kyle Brown and Bruce > Whitenack at Knowledge Systems Corp. The link I had to it is now > obsolete (it appears KSC is redoing their website) but if you can > contact one of the authors they might be able to make it available to > you. As of last November Kyle was still at KSC - try mailing him at > kbrown(a)ksccary.com to see if/how you can get a copy of the pattern > language. There were two parts to it the last time I knew - the "basic" > patterns and the "architectural" patterns. > > I also found a paper on an architecture called HORA which might be of > interest. You can find it at > http://object.idx.com/papers/horav2.pdf. > > I hope this helps. > > Bob Jarvis > The Timken Company > > >-----Original Message----- > >From: Mark A. Schwenk [SMTP:mas@wellthot.com] > >Sent: Wednesday, July 01, 1998 2:24 PM > >To: squeak(a)cs.uiuc.edu > >Subject: Re: Leads for database application development > > > >Jerome, > > > >Thanks for responding to my query. I am very interested in hearing your > >ideas and would be willing to work on this. It's long been a > goal of mine to > >produce a Smalltalk database application development environment. Your > >experience in the analysis of this domain sounds like it would be very > >helpful. > > > >Thanks! > > > >-Mark > > > >-----Original Message----- > >From: Jerome Garcia <Jerome.Garcia(a)wj.com> > >To: squeak(a)cs.uiuc.edu <squeak(a)cs.uiuc.edu> > >Date: Wednesday, July 01, 1998 12:31 PM > >Subject: Re: Leads for database application development > > > > > >> Mark, > >> > >> Bob Adler, Chris Earnest, and I were involved at Claris (Apple) in > >> development of a set of classes in C++ to provide distributed > >> heterogeneous database access. > >> > >> Bob designed a set of database provider classes which abstracted > >> access to various sources which could range from delimited > flat files > >> to the various relational databases. > >> > >> Chris designed a set of classes which could optimize distributed > >> heterogeneous requests. > >> > >> I designed a set of classes to represent the distributed data > >> dictionary and database requests. The database request classes were > >> based on but not limited to the ANSI SQL 92 standard. All database > >> request classes were subclasses of an expression class whose super > >> class was a tree node class. > >> > >> This was all part of a fairly ambitious effort at Claris. I left > >> shortly after proof of concept and initial testing of those parts > >> involving Bob, Chris, and I so that I could do Smalltalk > development > >> at Penergy. What can I say? I wanted to do Smalltalk :-) > >> > >> The parts which Bob and I did worked extremely well and > were very easy > >> to extend. I would like to see a similar direction taken > in Squeak. I > >> have thought about about doing this myself but have not > been able to > >> find the time. Also, I do not have any experience in > access of object > >> oriented databases so I can not say whether it is useful to include > >> them when considering the problem. > >> > >> If you would like additional info, send me some e-mail. > >> > >> Jerome E. Garica, Adventurous Mind > >> jegarcia(a)adventurousmind.com > >> > > >

1 0

Wow!!!
by Randy Siler 02 Jul '98

02 Jul '98

Wow!!! I just finished reading Scripting with Players -- The "User-Scripting" Style from the 2.1 release and all I can say is Wow!!! This is so exciting! It blows me away what you all are accomplishing. Thanks for all this great work, Randy Siler Coordinator of Computing, CIDR Center for Instructional Development & Research 396 Bagley Hall, BOX 351725 University of Washington Seattle, WA 98195-1725 Email: siler(a)cidr.washington.edu Phone: 206-543-6588 FAX:206-685-1213

1 0

Latest version of Squeak? (cont.)
by James O'Connor 02 Jul '98

02 Jul '98

All, Apparently I do have version 2.0 of Squeak for Windows95, which seems to be the latest blessed version. Take care Jay O'Connor joconnor(a)roadrunner.com "God Himself plays the bass strings first, when He tunes the soul"

1 0

Location of latest Squeak?
by James O'Connor 02 Jul '98

02 Jul '98

All, HI, I'm afraid I've fallen a little behind in keeping up to date with the latest version of Squeak. I think I'm using 1.2 or 1.23. I'm afraid I'm not quite sure where to FTP the latest verison of Squeak for Windows95. Could someone direct me to the location and needed files. Thank you, Take care, Jay O'Connor joconnor(a)roadrunner.com "God Himself plays the bass strings first, when He tunes the soul"

1 0

Re: Leads for database application development
by Jerome.Garcia＠wj.com 01 Jul '98

01 Jul '98

Mark, Bob Adler, Chris Earnest, and I were involved at Claris (Apple) in development of a set of classes in C++ to provide distributed heterogeneous database access. Bob designed a set of database provider classes which abstracted access to various sources which could range from delimited flat files to the various relational databases. Chris designed a set of classes which could optimize distributed heterogeneous requests. I designed a set of classes to represent the distributed data dictionary and database requests. The database request classes were based on but not limited to the ANSI SQL 92 standard. All database request classes were subclasses of an expression class whose super class was a tree node class. This was all part of a fairly ambitious effort at Claris. I left shortly after proof of concept and initial testing of those parts involving Bob, Chris, and I so that I could do Smalltalk development at Penergy. What can I say? I wanted to do Smalltalk :-) The parts which Bob and I did worked extremely well and were very easy to extend. I would like to see a similar direction taken in Squeak. I have thought about about doing this myself but have not been able to find the time. Also, I do not have any experience in access of object oriented databases so I can not say whether it is useful to include them when considering the problem. If you would like additional info, send me some e-mail. Jerome E. Garica, Adventurous Mind jegarcia(a)adventurousmind.com ______________________________ Reply Separator _________________________________ Subject: Leads for database application development Author: "Mark A. Schwenk" <mas(a)wellthot.com> at INTERNET Date: 6/30/98 8:08 PM I am looking at using Squeak for database application development (the sorts of projects I am currently using Microsoft Access for) and am wondering if anyone has leads on relevant current projects or has suggested directions of what Squeak needs in this area before I begin my work. I can't wait to start playing (ahem, working) with Squeak. -Mark Schwenk, WellThot Inc. mas(a)wellthot.com

2 1