Talk:Dave Cutler

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Untitled[edit]

See Talk:RSX-11 for my memories of Cutler. Note the puns. Cutler definitely intended WNT as a riff on VMS. Ortolan88

In Bruce Ellis's book "A Hitchhikers Guide to VMS", the evil-boss character is named David Cutlery. —Preceding unsigned comment added by 208.67.104.4 (talkcontribs) 18:57, December 21, 2006


I sold him my debugging and disassembling program for the LINC computer at his pre-DEC company for a cut of the profits (he offered me $200, which I stupidly refused). I believe not one copy was ever sold. My first demonstration of my ineptness at business. David 21:44 Nov 12, 2002 (UTC)


There's a biography of Cutler at http://213.169.196.120/HTML/publications/Bulletin/DHBUL%2067%20Nov-1995.pdf but it's in Dutch and I couldn't get anything but the birth date out of it. Ortolan88


I didn't have much luck looking for his bio on the Web, other than brief overviews starting in the 1990's. David 20:24 Nov 13, 2002 (UTC)


As far as I know, he left DEC and has been at Microsoft ever since. Was the NT on Alpha a joint project with MS that Compaq cancelled as part of the decades-long bemusement about what to do with Alpha? Ortolan88

Yes, pretty much. Jeh 04:19, 20 February 2007 (UTC)[reply]
Cutler 'left' Microsoft in the spring of 1996. Oh yes he's 'officially' on board but that's only because Microsoft don't want to lose his name. Instead they've helped finance Cutler's race car passion. As Cutler himself said of the deal: 'it keeps me from pissing all over them'. By the time of the Denver DC the word was getting out and Microsofties were whispering in panic 'Dave is gone! Dave is gone!' —Preceding unsigned comment added by 90.5.7.87 (talkcontribs) 20:47, August 31, 2008
That's an amusing claim. I recall being in his office hearing all about the multiple ready queue arrays that were going into Windows Server 2003. That was obviously quite a few years after 1996. Jeh (talk) 03:37, 11 September 2010 (UTC)[reply]

However great Cutler's wit, the Windows NT name was derived from 'NT OS/2', the previous name for the project. http://www.microsoft.com/windows2000/server/evaluation/news/fromms/kanoarchitect.asp —Preceding unsigned comment added by 68.43.142.209 (talkcontribs) 13:55, April 27, 2004

But there is still the wit. He's incessantly fascinated by Kubrick as the hardware abstraction layer testifies. Also there's the well known 'good morning Dave'. It's just a bunch of fun coincidences of course. —Preceding unsigned comment added by 90.5.7.87 (talkcontribs) 20:47, August 31, 2008

1) I think the "worked on RSX-11" is a little misleading, as I believe he only worked on the "M" version 2) Having spent years working on 11M, I remember his O/S code comments to be both funny and profane. —Preceding unsigned comment added by 74.93.101.81 (talkcontribs) 23:35, August 1, 2007

How delightful! —Preceding unsigned comment added by 90.5.7.87 (talkcontribs) 20:47, August 31, 2008

Wrong content?[edit]

I removed the background info on porting to Alpha because it was either wrong or very confusing. Sine Emerald and Prism are not VMS nor x86 projects, they cannot be about porting VMS to IA32 (x86). —Preceding unsigned comment added by 62.58.98.129 (talkcontribs) 12:06, January 5, 2005

Biography[edit]

If you're looking for a bio, there's a mini-auto-bio quoted from his book here. --Rebroad 21:31, 30 Mar 2005 (UTC)

Quotes[edit]

Someone should move the quotes to WikiQuotes. I'm not too familiar with said WikiMedia project, so I'd appreciate it if someone else did it. —Preceding unsigned comment added by Seec77 (talkcontribs) 16:46, June 10, 2005

The "fork queue" reference is pretty obscure. I assume it's completely meaningless for non technical people, and I admit that the meaning is not obvious to me either, even as a computer scientist. Can anyone explain this or clarify the reference? Timbatron 03:20, 23 October 2006 (UTC)[reply]

Fork queue is pronounced "fuck you". The fork list is a list of processes waiting to have access to the CPU, which is more or less the most important thing an OS does. Ortolan88 17:45, 19 November 2006 (UTC)[reply]
I worked for DEC for many years and never heard anyone pronounce it that way. btw, it has nothing to do with processes. Even though the things that were run that way were called "fork processes", they were really just procedure calls and no process context switches (or, in modern versions, thread context switches) are involved. NT calls the same concept a "deferred procedure call", which is much more evocative. Jeh 09:07, 2 May 2007 (UTC)[reply]
4Q2 is how I first learned this gag, in reference to some obscure command-control or telephony system. Orcmid 21:07, 26 November 2006 (UTC)[reply]

Windows Live[edit]

The article mentions he moved to Windows Live in 2006, this is contradicted by http://www.microsoft.com/presspass/exec/techfellow/Cutler/default.mspx which states he is "Currently responsible for the design of the 64-bit release of the Windows Operating System". I suggest we remove the live comment. RobChafer (talk) 09:59, 8 January 2008 (UTC)[reply]

I have personal information that supports the "moved to Windows Live" claim. Nothing I can provide as a reference atm, though. Jeh (talk) 10:58, 8 January 2008 (UTC)[reply]
Fair enough ... I just doubt it :) RobChafer (talk) 10:50, 14 February 2008 (UTC)[reply]
And sorry but I have first hand information that regardless of where he was officially moved he wasn't there anymore. He got sick and tired of them, did a deal to stay silent, and left. He turns up for dos and MS want his name around but he is definitely out of the picture by the spring of 1996. And that's first hand info - from the former developers on his team right outside the Tribe. —Preceding unsigned comment added by 90.5.7.87 (talkcontribs) 20:47, August 31, 2008
It seems he moved to Windows Azure (cloud computing). This was announced at the Azure launch and mentioned by Mark Russinovich in this Microsoft video (he calls the project 'red dog' which is the code name for Azure): http://channel9.msdn.com/shows/Going+Deep/Mark-Russinovich-Inside-Windows-7/ RobChafer (talk) 14:29, 28 January 2009 (UTC)[reply]
Haha. Yeah right. — Preceding unsigned comment added by 82.255.20.188 (talkcontribs) 01:22, 25 August 2015‎

Showstopper!(reference)[edit]

For those of you needing extra sources, a great book with lots of information is Showstopper! by G. Pascal Zachary. If I have time, I will begin using it to add info to the article. Emprovision (talk) 19:37, 27 May 2008 (UTC)[reply]

Yes it is a good book and it's required reading but watch your copyright infringements. Actually I'm surprised it wasn't read by the people who wrote this article. —Preceding unsigned comment added by 90.5.7.87 (talkcontribs) 20:47, August 31, 2008
That explains why the article is such crap. — Preceding unsigned comment added by 82.255.20.18801:20, 25 August 2015‎ (talkcontribs)
Yeah, I am currently in possession of the book, but I'm not great with entering information on Wikipedia. Anyone have any ideas on sections necessary for the article that I may be able to add? Emprovision (talk) 17:20, 21 April 2009 (UTC)[reply]

Why so little on original dev and so much on 64bit?[edit]

It seems strange that only one sentence is devoted to the development of the original NT project, then three to its development on 64bit platforms. On http://www.winsupersite.com/reviews/winserver2k3_gold1.asp it mentions the early versions of NT were built on the i860 and then the MIPS R3000 chip, and NT 3.51 was built around the PowerPC. From reading Showstopper a while back, I remember that his team's original build targets were not the x86. Cutler did this in order to ensure portability of the code, so then they would port to the x86 afterward for the commercial release. I thought perhaps NT4? was originally developed on the 64bit Alpha chip. So why did Cutler back away from lead development on 32bit to concentrate on 64bit platforms? How did his role at Microsoft change from the origin of Windows NT to his later development projects? Tumacama (talk) 17:33, 12 July 2009 (UTC)[reply]

The developers were in a panic in 1998. Revisiting Redmond, one saw they were all distraught. 'Where's Dave?' they asked. Dave had left right after NT 4.0 but got to keep his title (and Microsoft sponsorship of his racing) because, as he put it, 'their deal stops me from pissing all over them'. — Preceding unsigned comment added by 82.255.20.188 (talkcontribs) 01:18, 25 August 2015‎

Fork queue[edit]

Listen, folks, fork queue is not an example of sardonic humor, but it is completely reasonable tech speak. The RSX11M (and VMS) fork list is of course a fork queue by its FIFO mode of operation, to serialize access to non-reentrant system data base. As such it is a queue. Nothing funny about that. Next you cite QIO (queue i/o) as an example, what? 70.137.152.65 (talk) 03:23, 11 September 2010 (UTC)[reply]

I agree. I worked for DEC in the VMS heyday and nobody I encountered ever said "fork queue" with anything but a straight face. The name is perfectly logical, given the term "fork" (not the same as "forking a process" in Unix) and the fact that RSX and VMS use a FIFO queue to maintain the list. (Do Unix people ever think that "forking a process" means, well, doing the nasty to it?) In short I support this deletion. Jeh (talk) 03:28, 11 September 2010 (UTC)[reply]
Seriously doubtful everyone was as dry as you. — Preceding unsigned comment added by 82.255.20.188 (talkcontribs) 01:15, 25 August 2015‎
When I tried to include fork queue in the RSX glossary, it was kiboshed in favor of fork list because fork queue was forbidden in all documentation due to its vulgar connotation. You can find it here and there in the code, but not in any public place. Cutler wore a t-shirt with only the words fork queue on it at his going-away pig roast at DEC. People who "worked at DEC during the VMS heyday"= Johnny-come-latelies. This is from Ortolan88. 65.96.202.28 (talk) 21:20, 13 December 2010 (UTC)[reply]

In fact referring to fork processes as fork processes, not procedures, is also rational, as they run in a shortened context, which is held in the fork block. This shortened context points to entry PC, pointer to issuing SCB, pointer to issuing UCB of associated interrupt context. (in RSX11M, in VMS slightly more elaborate but same tune) This fork process is also not invoked as a subroutine, but as a coroutine, as RSX11M is in fact a nightmare from coroutine land in a lot of places and very tricky but clever with these tricks. This kind of coroutine linkage makes the execution also more process like. 70.137.152.65 (talk) 04:07, 11 September 2010 (UTC)[reply]

Sardonic Humour[edit]

"In addition to his engineering skills, Cutler is known for his sardonic humor. Sometimes even his error messages turn out to have a double meaning." - can we have a few examples or citation ?82.39.157.193 (talk) 21:41, 5 October 2010 (UTC)lanzarotemaps[reply]

IE.NFW -69 Path lost to partner "NFW" is an initialism for "no fucking way". 65.96.202.28 (talk) 21:14, 13 December 2010 (UTC)[reply]

The system is ripe with Kubrick references. Or don't you know that anymore? A lawyer stopped him in a corridor in Redmond in about 1990, two years after Dave's arrival, and told him he'd noticed that 'VMS', when advanced in the alphabet, became 'WNT', just as with the HAL Cutler references over and over, HAL of course (sigh) being the computer in 2001 and also the acronym for Cutler's hardware abstraction laywer. Cutler looked at the lawyer and responded: 'so it took you two years to figure that out?' — Preceding unsigned comment added by 82.255.20.188 (talkcontribs) 01:13, 25 August 2015‎

No Vista[edit]

I removed the "as well as Vista" sentence as this is not true.[citation needed] Windows Vista has been a major catastroph in Microsoft life[citation needed] precisely because Cutler retired shortly after the end of XP.[citation needed] Some rumors[citation needed] pretend[original research?] that Microsoft hired him back as a consultant to give guidelines for Windows Seven... We can all see the difference today.[original research?] --Casablanca1950 (talk) 11:50, 18 October 2010 (UTC)[reply]

Citation needed for the "retired shortly after" claim, for the "Vista was a catostrophe" claim, and for the "because" claim. Comments about living persons, even on talk pages, must comply with WP:BLP. Jeh (talk) 20:39, 18 October 2010 (UTC)[reply]

Agrippa-Ord Corporation[edit]

I found this address in a DECUS catalog PDF online, dated June, 1968:

N. David Culver, President
AGRIPPA-ORD Corporation
Monument Square
Carlisle, Massachusetts 01741

Note: There is no "Monument Square" in Carlisle, but there is a "Monument Square" in Concord. The "N" is given as a first name instead of second, and "Culver" is probably a misspelling.

Also, I found this patent reference online:

"Apparatus and method for synchronization of arithmetic exceptions in parallel pipelined execution units"
Singapore Patent 95904991
Inventor: N. DAVID CUTLER, BELLEVUE, WASHINGTON (and 4 others)
Applicant/Proprietor: DIGITAL EQUIPMENT CORPORATION

Putting these together hints strongly at what I stated years ago in this article, that David Cutler founded Agrippa-Ord, in Concord, MA. I can't actually find an independent reference showing this, though. Maybe we'll never know for sure until Cutler himself visits Wikipedia, or someone asks him. David Spector (talk) 04:15, 4 March 2011 (UTC)[reply]

Need more coverage of RSX-11M/M+[edit]

There should be more coverage of Cutler's key role in developing the RSX-11M opsys for the PDP-11. Not to be confused with more ungainly RSX-11D, the short-lived IAS opsys, nor the RSX-11S opsys, which was a stripped-down run-time-only version of RSX11-M. Cutler was a major factor in the "upstart" RSX-11M's ascendancy to supplant all the other versions (save -S), including IAS (which was supposed to be the flagship PDP-11 opsys) with the enhanced RSX-11M+. When Cutler quietly bowed out of RSX-11M development, rumors leaked out that he was heavily involved in what eventually became VAX/VMS.

Cutler can claim to to be a principal architect of the dominant 16-bit (RSX-11M), 32-bit (VAX/VMS), and Intel (Windows NT/2000/XP) commercial operating systems of their respective eras. To the extent that information can be found, his earliest major opsys accomplishment, RSX-11M, should be better documented, especially the concepts and techniques that influenced the development of later projects, both under Cutler and elsewhere. Reify-tech (talk) 23:30, 21 March 2011 (UTC)[reply]

While at duPont in the late 60's, Cutler first got heavily into operating systems studying Univac 1108 code on the midnight shift. He authored an operating system for the PDP10 which was used for data collection and analysis from a huge array of Instrons, and then an RSX-11A (it did support disk resident tasks) pre-cursor to RSX-11M. I was one of the very early users in the early 70's and wrote a number of device drivers for it and applied it in a variety of PDP-11 based process control systems before RSX-11M. It was a great day when he delivered a version supporting the PDP11 memory management module, so we could use the remaining 4K of the 32K physical memory and tasks could actually run protected from each other! The fork queue in RSX11A was the mechanism Cutler employed to achieve guaranteed interrupt response time, which was almost a religion with him. Non-interruptible processing was performed, the remaining processing placed on the fork queue to be continued as a co-routine of the interruptible operating system thread, and the interrupt dismissed. For us, RSX11A turned the PDP11 from a toy into a usable and inexpensive platform for manufacturing support in all sorts of applications until RSX-11M in '74. Cutler was indeed a prodigious producer of code. I spent numerous early mornings pouring thorough his books of listings which were like the Library of Congress. He was the first person I know to generate metrics on LOC/day. He went to our boss and demanded pay proportional to LOC/day. Not exactly the duPont culture of the day. At one point the DEC RSX-11D team visited us to share their plans, and Cutler, in his inimitible style, really took them to task about what real time computing was all about. Shortly after that he left for Digital armed with RSX-11A and his incredible talents. — Preceding unsigned comment added by Love2code (talkcontribs) 09:15, 21 April 2011 (UTC)[reply]
Thank you Love2code, for your very relevant comments. If you can find verifiable third-party documentation, these insights could be promoted to the main article. By the way, it is customary to append ~~~~ to comments on Talk pages, to "sign" one's remarks; otherwise the SineBot process will have to clean up for you. And welcome to Wikipedia editing! Reify-tech (talk) 15:48, 22 April 2011 (UTC)[reply]
Coverage of Cutler's key role in developing RSX-11M is still badly needed. As the article stood yesterday, nothing was mentioned about his role at Digital from his appearance in 1971 until he worked on VMS in 1974. This is very misleading, since his key role in RSX-11M is the very reason he was tapped to lead the VMS effort. Much of the internal structure, concepts, terminology, and APIs developed for RSX-11M were migrated directly into VMS, and later into Windows NT. I don't have access to the old RSX-11M documentation anymore, but somebody out there must have followed the successive development of the three operating systems and documented the continuity of their development. Reify-tech (talk) 17:04, 10 June 2017 (UTC)[reply]

Merge sections[edit]

I added a section regarding Cutler's current work on the Xbox team. Re-reading the article, it seems like both Windows Azure and Xbox should be combined in to a single section, as they are both rather short blurbs. Perhaps a title of 'Post-NT projects', 'Post-NT work' or 'Later work within Microsoft'? A new section seemed as bit much, without first asking for feedback. However, if no one raises any issues in the next few days/weeks, I will likely merge the two. agentlame (talk) 08:05, 3 February 2012 (UTC)[reply]

As a person[edit]

What I don't get when I read this article is how he is as a person.

I know some things from other sources (he hated Unix for good reasons, he's an aggressive asshole) but I don't know much detail. — Preceding unsigned comment added by 82.139.81.0 (talkcontribs) 20:59, 18 March 2013‎

Impertinence seems to be your strong suit. As a person, he's the kind of guy who would either step over or on you. Good? — Preceding unsigned comment added by 82.255.20.188 (talkcontribs) 01:08, 25 August 2015
Arguably the reason he hated Unix so much is he was one of the principal engineers behind VMS, an OS that aimed to and ultimately failed at "killing" Unix. I submit if one were to ask him why Unix is so inferior he couldn't name a single real technical reason that's relevant today (His "get a byte" song seems ignorant of the fact Unix can do block I/O.). — Preceding unsigned comment added by 146.145.85.49 (talk) 01:22, 23 May 2014 (UTC)[reply]
No not at all. He used to say that Unix was written by a 'committee of PhDs'. What he didn't like was that too many people were involved in its development down the line. He's said on more than on occasion that an OS should be the brainchild of single system architect. — Preceding unsigned comment added by 82.255.20.188 (talkcontribs) 01:08, 25 August 2015
Did you all know that WP:BLP applies to talk pages too? From the lede: "This policy applies to any living person mentioned in a BLP, whether or not that person is the subject of the article, and to material about living persons in other articles and on other pages, including talk pages." (emph. added) Jeh (talk) 00:20, 11 June 2017 (UTC)[reply]