BABYL OPTIONS: -*- rmail -*- Version: 5 Labels: Note: This is the header of an rmail file. Note: If you are seeing it in rmail, Note: it means the file has no messages in it.  1,, Mail-from: From strang@HEISENBERG.MIT.EDU Tue Oct 20 16:43:52 1998 Return-Path: Received: from MIT.EDU (PACIFIC-CARRIER-ANNEX.MIT.EDU) by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA08875; Tue, 20 Oct 98 20:44:11 EDT Received: from HEISENBERG.MIT.EDU by MIT.EDU with SMTP id AA09200; Tue, 20 Oct 98 20:44:08 EDT Received: (from strang@localhost) by HEISENBERG.MIT.EDU (8.8.5/8.8.5) id UAA12169; Tue, 20 Oct 1998 20:43:52 -0400 From: Ricardo J Torres Message-Id: <199810210043.UAA12169@HEISENBERG.MIT.EDU> To: 6044-forum@MIT.EDU Cc: strang@MIT.EDU Subject: Instruction Set for register machine Date: Tue, 20 Oct 1998 20:43:52 EDT *** EOOH *** From: Ricardo J Torres To: 6044-forum@MIT.EDU Cc: strang@MIT.EDU Subject: Instruction Set for register machine Date: Tue, 20 Oct 1998 20:43:52 EDT I'm fairly certain we need to add a SET0 (MEM(0) <- 0) instruction to my instruction set. The reason is that with the current instruction set we can't set the contents of register 0 to an abitrary number without doing a branch. The reason the current system doesn't work is because currently setting register 0 to 0 requires a branch on 0 operation. Branch on 0 has no address arguments, it requires the target branch location (if register 0 == 0) to be in register 1. In order to be able to set the target address in register 1 you have to set it in register 0 (and then use a copy0to1). This requires setting register 0 to 0, and then incrementing register 0 to the target address. This is circular, as setting register 0 to 0 is the reason we are branching in the first place. A SET0 instruction fixes this (and I think makes things meet the specs). I would imagine a SET0 operation would be very economical, I am however sorry I didn't see this in class. Ricardo  1,, Mail-from: From meyer@lcs.mit.edu Tue Oct 27 23:41:02 1998 Return-Path: Received: from shorthair.ne.mediaone.net (greifmeyer.ne.mediaone.net) by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA20392; Tue, 27 Oct 98 23:41:00 EST Message-Id: <000101be022d$236bbe00$229e8018@shorthair.ne.mediaone.net> From: "Albert R. Meyer" To: <6044-forum@theory.lcs.mit.edu> Subject: Assignment and BNF on web Date: Tue, 27 Oct 1998 23:40:48 -0500 Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-Msmail-Priority: Normal X-Mailer: Microsoft Outlook Express 4.72.3110.5 X-Mimeole: Produced By Microsoft MimeOLE V4.72.3155.0 *** EOOH *** From: "Albert R. Meyer" To: <6044-forum@theory.lcs.mit.edu> Subject: Assignment and BNF on web Date: Tue, 27 Oct 1998 23:40:48 -0500 X-Priority: 3 X-Msmail-Priority: Normal X-Mailer: Microsoft Outlook Express 4.72.3110.5 X-Mimeole: Produced By Microsoft MimeOLE V4.72.3155.0 Take a look at the course web page: http://theory.lcs.mit.edu/classes/6.044/ it has the assignment due Monday and some fresh notes on Scheme grammar. Regards, A.  1,, Mail-from: From meyer@theory.lcs.mit.edu Tue Oct 27 18:44:23 1998 Return-Path: Received: from stork.lcs.mit.edu by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA20416; Tue, 27 Oct 98 23:44:30 EST Received: by stork.lcs.mit.edu (5.65c/TOC-1.2C) id AA08766; Tue, 27 Oct 98 23:44:23 EST Date: Tue, 27 Oct 98 23:44:23 EST Message-Id: <199810280444.AA08766@stork.lcs.mit.edu> From: "Albert R. Meyer" To: 6044-forum@theory.lcs.mit.edu Subject: test *** EOOH *** Date: Tue, 27 Oct 98 23:44:23 EST From: "Albert R. Meyer" To: 6044-forum@theory.lcs.mit.edu Subject: test ignore this  1,, Mail-from: From pnkfelix@MIT.EDU Tue Oct 27 22:13:12 1998 Return-Path: Received: from MIT.EDU (SOUTH-STATION-ANNEX.MIT.EDU) by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA21768; Wed, 28 Oct 98 03:13:20 EST Received: from MINT-SQUARE.MIT.EDU by MIT.EDU with SMTP id AA23101; Wed, 28 Oct 98 03:13:09 EST Received: by mint-square.mit.edu (8.8.7/4.7) id DAA22954; Wed, 28 Oct 1998 03:13:12 -0500 (EST) Message-Id: <199810280813.DAA22954@mint-square.mit.edu> To: "Albert R. Meyer" Cc: 6044-forum@theory.lcs.mit.edu Subject: Re: test In-Reply-To: Your message of "Tue, 27 Oct 1998 23:44:23 EST." <199810280444.AA08766@stork.lcs.mit.edu> Date: Wed, 28 Oct 1998 03:13:12 EST From: Felix S Klock *** EOOH *** To: "Albert R. Meyer" Cc: 6044-forum@theory.lcs.mit.edu Subject: Re: test In-Reply-To: Your message of "Tue, 27 Oct 1998 23:44:23 EST." <199810280444.AA08766@stork.lcs.mit.edu> Date: Wed, 28 Oct 1998 03:13:12 EST From: Felix S Klock Prof. Meyer et. all- I was looking over R5RS in preperation to do some hacking on this problem set and noted that it defined (begin ... ) as a library syntax; IE something that can be defined in terms of the other functions. In lecture you said that begin was one of the necessary parts of our kernel. After giving it some thought I came to this possible desugaring of begin in terms of lambda: (begin ) => ((lambda (x) ((lambda (x) ) )) ) Since arguments to a lambda expression are evaluated before the lambda body is evaluated, doesn't this ensure that expressions 1 through 3 will be applied in the correct sequence? And since the argument to each lambda expression isn't used in the body of the lambda, that ensures that the value of the expression as a whole is guaranteed to be the value of evaluating the inner most expression in the global environment (which could have been altered by the preceding evaluations of expressions 1 through 2?) A more general form for this desugaring would be (begin ... ) ((lambda (x) ((lambda (x) ((lambda (x) ; ... N-1 of these "((lambda (x)" ((lambda (x) ((lambda (x) ((lambda (x) ) )) )) )) ; ... expression N-1 through 2 in series with "))" appended )) )) ) However, I am also aware that R5RS defines the body of a lambda expression to be able to take multiple expressions, so they could have just meant that (begin ) desugars to ((lambda () )) Please let me know if the above desugaring for kernel-scheme is invalid. I think the general form I listed may fail on single expression begins (begin ) But perhaps someone else can think of a clever way to express the desugaring macro that doesn't have a problem with the single expr. (I also could be wrong about it failing...since I didn't define the macro very formally, its hard to tell how it will handle that case) Just thought it might be a interesting way to make your kernel-scheme even smaller. Also I noted that R5RS says that letrec is also a library syntax. I couldn't figure out how the writers of R5RS thought it could be desugared though. -Felix  1,, Mail-from: From meyer@theory.lcs.mit.edu Wed Oct 28 06:25:08 1998 Return-Path: Received: from stork.lcs.mit.edu by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA26401; Wed, 28 Oct 98 11:25:16 EST Received: by stork.lcs.mit.edu (5.65c/TOC-1.2C) id AA08963; Wed, 28 Oct 98 11:25:08 EST Date: Wed, 28 Oct 98 11:25:08 EST Message-Id: <199810281625.AA08963@stork.lcs.mit.edu> From: "Albert R. Meyer" To: pnkfelix@MIT.EDU Cc: 6044-forum In-Reply-To: <199810280813.DAA22954@mint-square.mit.edu> (message from Felix S Klock on Wed, 28 Oct 1998 03:13:12 EST) Subject: Re: test Reply-To: 6044-lecturer@theory.lcs.mit.edu *** EOOH *** Date: Wed, 28 Oct 98 11:25:08 EST From: "Albert R. Meyer" To: pnkfelix@MIT.EDU Cc: 6044-forum In-Reply-To: <199810280813.DAA22954@mint-square.mit.edu> (message from Felix S Klock on Wed, 28 Oct 1998 03:13:12 EST) Subject: Re: test Reply-To: 6044-lecturer@theory.lcs.mit.edu Yes, our "kernel" is not minimal. In designing a "substitution model," I wanted a user to be able to watch her code as it ran. If I desugared too much, the running code would look really alien, so I chose a modest, but not minimal kernel. BTW, your desugaring of BEGIN looks about right, except for the problem that you put expressions in the scope of a LAMBDA (X) which can cause trouble if they have a free X. Here's the general desugaring which avboids this problem: (make-simple-rule '(begin ?expr) '?expr) (make-simple-rule '(begin ?expr1 ?expr2 ~exprs) '(let ((v1 ?expr1)) (begin ?expr2 ~exprs))) Remember that LET can be treated as sugar: (make-general-rule '(let (~binds) ?expr) (lambda (binds expr) `((lambda (,(variables-of-binds binds)) ,expr) ,@(expressions-of-binds binds)))) BTW, desugaring (BEGIN ~stuff) into LAMDBA () ~STUFF) just postpones the problem, since lam-exps with bodies consisting of a sequence of exprs, as opposed to a single expr are not in the kernel and would have to be desugared. Regards, A.  0, unseen,, *** EOOH *** Mail-from: From meyer@theory.lcs.mit.edu Wed Oct 28 06:31:16 1998 Return-Path: Received: from stork.lcs.mit.edu by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA26508; Wed, 28 Oct 98 11:31:24 EST Received: by stork.lcs.mit.edu (5.65c/TOC-1.2C) id AA08985; Wed, 28 Oct 98 11:31:16 EST Date: Wed, 28 Oct 98 11:31:16 EST Message-Id: <199810281631.AA08985@stork.lcs.mit.edu> From: "Albert R. Meyer" To: 6044-forum@theory.lcs.mit.edu In-Reply-To: <199810281625.AA08963@stork.lcs.mit.edu> (meyer@theory.lcs.mit.edu) Subject: desugaring BEGIN: correction CORRECTION: (make-simple-rule '(begin ?expr) '?expr) (make-simple-rule '(begin ?expr1 ?expr2 ~exprs) '(let ((v1 ?expr1) (v2 (lambda () (begin ?expr2 ~exprs))) (v2)))) (My earlier version put EXPR2 and ~EXPRS in the scope of V1, which can cause problems if they have a free occurrence of V1.) Regards, A.  0, unseen,, *** EOOH *** Mail-from: From strang@HEISENBERG.MIT.EDU Thu Oct 29 18:09:27 1998 Return-Path: Received: from MIT.EDU (SOUTH-STATION-ANNEX.MIT.EDU) by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA21707; Thu, 29 Oct 98 23:09:48 EST Received: from HEISENBERG.MIT.EDU by MIT.EDU with SMTP id AA20267; Thu, 29 Oct 98 23:09:34 EST Received: (from strang@localhost) by HEISENBERG.MIT.EDU (8.8.5/8.8.5) id XAA09257; Thu, 29 Oct 1998 23:09:28 -0500 From: Ricardo J Torres Message-Id: <199810300409.XAA09257@HEISENBERG.MIT.EDU> To: 6044-forum@MIT.EDU Subject: zephyr instance Date: Thu, 29 Oct 1998 23:09:27 EST If people are using athena for the problem set due Monday (or any other project), they may want to try the 6.044 zephyr instance. This would allow instantaneous feed back to others listening, since there is no need to incorporate mail. The down side is, you have to be on athena and and you may always feel pressured to keep up your side of the conversation even if you are busy doing something else. In orer to receive 6.044 zephyrs, you have to be running zwgc (most athena setups do this by default) and be subscribed to the instance. SUBSCRIBING: To subscribe on athena, each time you login (or in an environment file) type: zctl sub message 6.044 \* If you don't want to type it each time you login (or in an environment file) and want to ADD it to your subscriptions until you delete it, type: zctl add message 6.044 \* To delete type 'zctl del message 6.044 \*' (I believe). Your subscription will remain active from one athena session to the next until you delete your subscription. SENDING: Now to send a a zephyr, type: zwrite -i 6.044 type message followed by line with just a period on it as usual. If you use any GUI zephyr tools, your mileage will vary. ZEPHYR LOG: THE ZEPHYRS WILL BE LOGGED to a file in the 'zlog' locker. You can get to this from any athena machine by typing: attach zlog The file holding the contents of the log is /mit/zlog/6.044 and is in plain text format. You can also get to the logs via nfs at 'sweet-transvestite.mit.edu:/u1/zlog/6.044' and from machines that support nfs automount, you can normally find this at '/net/sweet-transvestite.mit.edu//u1/zlog/6.044.' I don't know if this will be useful, but I figured I'd give it a shot. Ricardo  1,, Mail-from: From tzumainn@MIT.EDU Sat Oct 31 16:58:02 1998 Return-Path: Received: from MIT.EDU (PACIFIC-CARRIER-ANNEX.MIT.EDU) by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA13951; Sat, 31 Oct 98 21:59:22 EST Received: from MIT.MIT.EDU by MIT.EDU with SMTP id AA19797; Sat, 31 Oct 98 21:59:13 EST Received: from ISHMAEL.MIT.EDU by MIT.MIT.EDU (5.61/4.7) id AA16565; Sat, 31 Oct 98 21:59:10 EST Message-Id: <3.0.2.32.19981031215802.006efb80@po9.mit.edu> X-Sender: tzumainn@po9.mit.edu X-Mailer: QUALCOMM Windows Eudora Light Version 3.0.2 (32) Date: Sat, 31 Oct 1998 21:58:02 -0500 To: 6044-forum@MIT.EDU From: Tzu-Mainn Chen Subject: Re: ps4 due friday Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" *** EOOH *** X-Sender: tzumainn@po9.mit.edu X-Mailer: QUALCOMM Windows Eudora Light Version 3.0.2 (32) Date: Sat, 31 Oct 1998 21:58:02 -0500 To: 6044-forum@MIT.EDU From: Tzu-Mainn Chen Subject: Re: ps4 due friday Hi, Professor Meyer sent me these suggestions for the first two problems of the problem set, and I thought that I'd pass them on. . . Tzu-Mainn Chen pa>Date: Fri, 30 Oct 98 22:07:20 EST >From: "Albert R. Meyer" >To: tzumainn@MIT.EDU >Subject: Re: ps4 due friday >Reply-To: 6044-lecturer@theory.lcs.mit.edu > >Concentrate on the one-variable case. A degree n polynomial goes to >+-infinity at an O(x^n) rate, so its range must be denser than powers of 2 >allow. The other, divisible-by-k, problem is solved by appealing to >arithmetic mod k. > >Hope that helps (without undermining the problems entirely). > >Perhaps we should broadcast this to the forum. What do you think? > >Regards, A. > > > X-Sender: tzumainn@po9.mit.edu > X-Mailer: QUALCOMM Windows Eudora Light Version 3.0.2 (32) > Date: Fri, 30 Oct 1998 04:07:08 -0500 > From: Tzu-Mainn Chen > > Hi, > I've been working on the first two problems of the problem set pretty much > all night now, and I still can't get a grip on either of them. I think the > problem is that I don't even know how to begin to approach either problem. > . . are there any hints that you could give. . . ? > > Tzu-Mainn Chen >  1,, Mail-from: From pnkfelix@MIT.EDU Sat Oct 31 18:01:36 1998 Return-Path: Received: from MIT.EDU (PACIFIC-CARRIER-ANNEX.MIT.EDU) by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA14156; Sat, 31 Oct 98 23:01:47 EST Received: from MINT-SQUARE.MIT.EDU by MIT.EDU with SMTP id AA24430; Sat, 31 Oct 98 23:01:38 EST Received: by mint-square.mit.edu (8.8.7/4.7) id XAA22603; Sat, 31 Oct 1998 23:01:37 -0500 (EST) Message-Id: <199811010401.XAA22603@mint-square.mit.edu> To: 6044-forum@MIT.EDU Subject: Weird... Date: Sat, 31 Oct 1998 23:01:36 EST From: Felix S Klock *** EOOH *** To: 6044-forum@MIT.EDU Subject: Weird... Date: Sat, 31 Oct 1998 23:01:36 EST From: Felix S Klock Hurm! Using big-O-notation to indicate density of the function...WEIRD! Never seen that method before. This may be a trivial proof, but certainly not elementary... Do any of you know if this method covered in 6.042? When I took 6.042 (fall of '97) I don't believe this method was ever covered. But I could be wrong. In any case, I think that gives me enough info to figure out the first problem, which was the real roadblock for me Okay, enough musings from me. -Felix From meyer@theory.lcs.mit.edu Wed Nov 4 11:38:29 1998 Return-Path: Received: from stork.lcs.mit.edu by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA06297; Wed, 04 Nov 98 16:38:41 EST Received: by stork.lcs.mit.edu (5.65c/TOC-1.2C) id AA14598; Wed, 04 Nov 98 16:38:29 EST Date: Wed, 04 Nov 98 16:38:29 EST Message-Id: <199811042138.AA14598@stork.lcs.mit.edu> From: "Albert R. Meyer" To: 6044-forum Subject: ps6 due Friday 11/13 is on the web Reply-To: 6044-lecturer@theory.lcs.mit.edu one submodel programming problem, two problems using the Submodel to explain Scheme behavior under renaming bound variables, and one problem on polynomial ranges. Regards, A. From meyer@theory.lcs.mit.edu Thu Nov 5 09:12:02 1998 Return-Path: Received: from stork.lcs.mit.edu by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA18755; Thu, 05 Nov 98 14:12:15 EST Received: by stork.lcs.mit.edu (5.65c/TOC-1.2C) id AA15276; Thu, 05 Nov 98 14:12:02 EST Date: Thu, 05 Nov 98 14:12:02 EST Message-Id: <199811051912.AA15276@stork.lcs.mit.edu> From: "Albert R. Meyer" To: 6044-forum Subject: submodel example Reply-To: 6044-lecturer@theory.lcs.mit.edu I hope at this point you have a basic understanding of how the Scheme submodel works. Below is a sample series of expressions and definitions WHICH I WOULD EXPECT YOU TO BE ABLE TO SIMULATE BY HAND in a few minutes, using SM.SCM as a reference. To be sure we're all together on this, on MONDAY I will give a 20 min. IN-CLASS QUIZ asking you to simulate a similar (slightly shorter) example. Regards, A. (define (make-account balance) (lambda (deposit) (begin (set! balance (+ balance deposit)) balance))) (define acct1 (make-account 10)) (acct1 0) (define acct2 (make-account 20)) (acct2 0) (show-submodel-environment) From pnkfelix@MIT.EDU Fri Nov 6 00:44:33 1998 Return-Path: Received: from MIT.EDU (SOUTH-STATION-ANNEX.MIT.EDU) by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA26213; Fri, 06 Nov 98 05:44:46 EST Received: from M37-312-23.MIT.EDU by MIT.EDU with SMTP id AA01836; Fri, 6 Nov 98 05:44:30 EST Received: by m37-312-23.mit.edu (SMI-8.6/4.7) id FAA23808; Fri, 6 Nov 1998 05:44:33 -0500 Message-Id: <199811061044.FAA23808@m37-312-23.mit.edu> To: 6044-forum@MIT.EDU Subject: SM interpreter Date: Fri, 06 Nov 1998 05:44:33 EST From: Felix S Klock Prof. Meyer (et al.)- Something in my gut is still reacting to the fact that the SM model is allowing different procedures to be eq? (different in the sense that they are bound to different variable, not different in the sense that they evaluate to differing arguments given some input)...so, in an effort to make myself more comfortable with the idea, I began doing some research on what R5RS says on the matter, and the reasons for why it says that... My reading of the rrrs-authors forum (and R5RS) uncovered something interesting... The authors, after much debacle about eq? and what it does (see http://www-swiss.ai.mit.edu/ftpdir/scheme-mail/HTML/rrrs-1986/threads.html for details) seemed to settle on standardizing what scheme does using the procedure eqv?, and leave a lot of eq? unspecified (for efficiency in individual impementations) Unfortunately, SM doesn't implement eqv? Also, eq? is supposed to behave identically to eqv? when applied to procedures. Now, as it turns out, the behavior of eqv? on examples such as (eqv? (lambda (x) x) (lambda (y) y)) is unspecified. Therefore in regards to its behavior on eq?, the Sub.Model is R5RS-compliant. So the only real job remaining is to implement eqv? ...which may just be a matter of binding it to eq? I'm not sure yet... I would like to note that in the arguments on rrrs-authors about eq?, there was support for requiring the above example to return false. So I feel justified for having that feeling in my gut. -Felix From strang@HEISENBERG.MIT.EDU Thu Nov 19 08:15:14 1998 Return-Path: Received: from MIT.EDU (SOUTH-STATION-ANNEX.MIT.EDU) by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA02546; Thu, 19 Nov 98 18:25:18 EST Received: from HEISENBERG.MIT.EDU by MIT.EDU with SMTP id AA13037; Thu, 19 Nov 98 13:16:51 EST Received: (from strang@localhost) by HEISENBERG.MIT.EDU (8.8.5/8.8.5) id NAA12566; Thu, 19 Nov 1998 13:15:14 -0500 From: Ricardo J Torres Message-Id: <199811191815.NAA12566@HEISENBERG.MIT.EDU> To: 6044-forum@MIT.EDU Cc: strang@MIT.EDU Subject: Web server Date: Thu, 19 Nov 1998 13:15:14 EST Looks like the web server is down again. Ricardo From meyer@theory.lcs.mit.edu Mon Nov 30 22:01:27 1998 Return-Path: Received: from stork.lcs.mit.edu by theory.lcs.mit.edu (5.65c/TOC-1.2S) id AA03079; Tue, 01 Dec 98 03:01:52 EST Received: by stork.lcs.mit.edu (5.65c/TOC-1.2C) id AA28839; Tue, 01 Dec 98 03:01:27 EST Date: Tue, 01 Dec 98 03:01:27 EST Message-Id: <199812010801.AA28839@stork.lcs.mit.edu> From: "Albert R. Meyer" To: 6044-forum Subject: problem 5 Reply-To: 6044-lecturer@theory.lcs.mit.edu Problem 5 has been added to the Handout 9 on Scheme Computability. Problems 2,3,4,5 are due Friday. There won't be a final exam. We'll have an in-class quiz next Monday; last class is the following Wed. Regards, A.