Sitewide
RSS Feed:
Why oh why oh why oh (RSDC Fytte 3)
 |
By: David Norfolk, Practice Leader - Development, Bloor Research Published: 10th June 2008 Copyright Bloor Research © 2008 |
 |
Delicious
Digg
reddit
Facebook
StumbleUponSo, RSDC has this team of hero programmers fighting the Dark Side of programming—using the Sword of Jazz and the Shield—or hiding behind something looking like a Second Life avatar.....
It's all so cool—but the programmers I meet in real life are mostly fairly serious, sometimes be-suited, professionals who get a kick out of doing a professional job, professionally. Why would they want to play at being Justice League of America? And are people who don't see themselves as comic book superheroes—the "techies in suits" I tried to cater for in Reg Developer (which now seems to have merged back into The Register)—in danger of becoming disenfranchised?
For some reason, I'm thinking of that old DnD cartoon (from White Dwarf? I do have a misspent youth, but a poor memory)—a group of elves and dwarfs are sitting around a table and one says "let's role play being a couple of programmers, an admin assistant and a computer op in a bedsit in Surbiton"—or something like that.
There is, of course, always room for someone like Grady Booch, who isn't noticeably be-suited (although he too could be role-playing these days). He's inspiring. His interest in "beauty in programming" is enlightening—perhaps we need beauty metrics to complement complexity metrics (and, boy, do we need to minimise complexity and maximise beauty in many of our current systems). And, his concern for Software Archaeology is timely—some of the greatest artefacts of the 20th Century are in danger of being lost forever (he's excavated the OS/370 source code, but OS/360 seems lost in the mud—even Fred Brooks doesn't have a copy). But he's a lot brighter and more useful than Green Lantern—and I really don't see why IBM had to play the geeky superhero card so heavily at RSDC.
However, Grady is keen on Second Life (which is about as un-serious and un-suited as you can get) as a home for "serious" working communities in the world of globalised development. He can meet a lot more people, more cheaply, on Second Life. And there was even talk of IBM and Linden Labs working on an Open Source community for Second Life. It is popular and successful and people are making real money in this virtual world....
I've had a virtual community (on CIX) at the centre of my working life since before I became a journalist—working for a Bank in the process of blood-on-the-carpet "Business Process Redesign" I needed to talk to people outside of my current "Business Process", without always making my real identity obvious. So, almost on principle, I love the idea of communicating with fellow developers in other environments, countries etc in a Virtual World.
Nevertheless, I've never found modern online communities to work very effectively up to now (CIX is an offline Forum-style thing) - partly because the current fad for "instant communication" means that you get instant information from busy people who are only half-thinking about your questions—crap (or at least incomplete or out-of-context) information at the speed of thought. And partly because the UI is often very clunky and unable to cope with Web latency.
Perhaps Second Life is better—now that I've actually got Broadband (a month after I first asked BT to switch over from ISDN) I may find out (at least I can now download the 35 megabytes of client software upgrade I now apparently need in a sensible time). However, despite Second Life's obvious success, it does occur to me that finding a million people on the Net to use something new and different, just because it is new and different, wouldn't be hard. My personal metric for the success of virtual community environments, BTW, is when I'm no longer expected to carry little squares of cardboard around with me, with my name and contact details on them
I did make a straw poll on serious Second Life usage at an Analyst Breakfast at RSDC at which some users were talking to us but this wasn't terribly conclusive. This was partly because 45% Second Life usage on a sample of 3 isn't terribly statistically significant; and partly because I mostly couldn't hear what anyone was saying over the noise of breakfast anyway. At least that's a problem I wouldn't have if we'd met in Second Life; but I doubt if the Eggs Benedict are very good there.
So, I'm left wondering about Second Life and virtual communities as a serious developer's tool—I'm not quite convinced yet. IBM is investing serious money in its own virtual island, I believe—now I need to actually meet some of the serious developers that are supposed to live on it. I hope to get my Second Life account re-invigorated RSN—but if any of them still read legacy blogs like this, perhaps they could drop me a line!
Blog: IBMRSDC08
Twitter Hashtag: #RSDC
Reader Comments
We are no longer accepting comments against this item. We suggest contacting the author directly.
10th June 2008: 'William Blair' said:
Sadly, you have been misinformed, or else did not understand what you were told.
> he's excavated the OS/370 source code
There is no such thing as OS/370. After OS/360 came OS/VS1 (for small System/370s) and OS/VS2 (for large System/370s). That could be what you meant by "OS/370." However, I actually suspect that you meant "MVS/370" which was the euphemism for OS/VS2 starting with Release 2.0 up through 3.8 (the last).
> but OS/360 seems lost in the mud—
> even Fred Brooks doesn't have a copy
Not so. I have a copy of all of OS/360 source, and I bet there are a thousand or more copies out there. Please Google "Hercules and OS/360." I also have a copy of MVS/370 3.8 source (along with a number of its so-called "Selectable Units"). There was never any need to "find" any of this material. IBM distributed it widely. It was never "lost." This fellow you are referring to, as well as you, are not in touch with what is actually going on in this world.
10th June 2008: 'David Norfolk' said:
Fair cop guv - but a Blog is (or should be) top-of-the- head conversation, not a researched article. NALOPKT.
You are, of course, right - I meant MVS/370. And I even worked (as a trainee than as an assembler programmer and IMS DBA) on the 360 and on the 370, so I did once remember all those names. But it was long ago.
I probably misheard Grady, but he was just chatting too. And I think that the essential point stands - we should be preserving this stuff in official collections, not just at the back of peoples' desks.
Anyway, can you email me your contact details, so I can pass them on to Grady, if he is looking for OS/360?
Thanks.
12th June 2008: 'Grady Booch' said:
well, given that, as you say, i'm not in touch with what is going on in the world, i'm not sure my reply will have any meaning to you, but i'll try anyway. :-)
> There is no such thing as OS/370.
you are correct. however, when i went digging in ibm's archives, i came across a product line item named "os/370" - an internal designation, i would think - that i was told represented the source for the os used on the 370. at any rate, ibm has in its archives this code.
however, when we had our near death experience in the late 90s, things were tossed off the sinking ship to help it float. apparently, the original 360 source was caught up in that.
> Not so. I have a copy of all of OS/360 source
and i'd love to obtain a copy, for preservation at the museum. however, i suspect that you have a later version of the os, and i was seeking the original code base - i can't find that anywhere, and neither fred nor ibm have a copy of that. what version of the os do you have the source for?
> IBM distributed it widely. It was never "lost."
depends on what you mean by lost, i suppose. the white house "lost" several months of emails, didn't they? :-) at any rate, the reality is that, as far as ibm's archivists tell me, they don't have a copy, and neither does the museum.
> This fellow you are referring to, as well as you, are not in touch with what is actually going on in this world.
well then, i'll suppose i'd best slink back away to my cave where i can continue to hide away from the real world...
30th July 2008: 'William Blair' said:
Grady asked:
> and i'd love to obtain a copy, for preservation at the museum.
Send me your address and I will send you a CD-ROM of the optional material, as well as the original installation MRM.
> however, i suspect that you have a later version of the os,
> what version of the os do you have the source for?
The very last, 21.8F.
> i was seeking the original code base > i can't find that anywhere,
Would you mean, by "original code base" Release 1 of OS/360 from Fall 1965?
If so, I know for a fact that it is lost to history. When IBM internally converted to the "CLEAR" source code management system, all the _cards_ and _tapes_ were thrown away. That was well after Release 1. CLEAR underwent two major revisions since that time, and was put to sleep more than a decade ago, so even if its databases could be found somewhere, it would not yield OS/360 Release 1, and probably not even OS/360 at all.
For amusement purposes, I too would love to have OS/360 Release 1 source (or any early release ... 4 was the first that ran worth a hoot, so I'd probably prefer that, anyway). But as I remember, it took a while before IBM even started distributing the "optional material" (i.e., source) on tapes. Few folks wanted it, because PTFs, which appeared in abundance even before a release became available, changed the source for the (ahem) "more interesting" modules. It was not at all unusual to discover at a SHARE meeting that dozens and dozens (literally) of other folks had a microfiche reader brought to a keypuch operator who then keypunched the source from the fiche listing of a PTF. Folks started sharing IBM OS/360 source code this way, and the first elements (not already distributed in source by IBM) that were of interest me I obtained this way.
Regardless, I predict that if the actual truth could be known somehow, there does not remain any machine readable version of any early OS/360 release. There are of course, boatloads of copies of the latest several releases, but the one that has generated the most widespread sharing is the very last one, which was the basis of maintaining it for many many years for folks that could not run OS/VS1 or OS/VS2 after IBM had moved on.
I'll be happy to send you (or anyone) a CD-ROM of the source (for 21.8), but it's long been available online for download from at least four sites that can be found using Google.
11th June 2008: 'David Wright' said:
Please, not more talk about beauty in software. "Beauty is in the eye of the Beholder..."; we don't need aesthetically pleasing software, we need well-designed software. Programmers are not artists, at best they are craftsmen who should aspire to be engineers. Yes, virtually anything can be seen as beautiful by someone, but that does not make it useful or easy to change. ...I have had this discussion a few times, and it always splits 50-50, artistry versus engineering; no wonder s/w projects have troubles.
12th June 2008: 'David Norfolk' (Author) said:
My view of "beauty" (and not just in software) comes from Morris and the Arts & Crafts movement - there is beauty in function. Beauty in software includes "fitness for purpose" and maintainability.
A well-made spanner has beauty of form and function; and one that bends when you use it can hardly be seen as a beautiful spanner (although it might have some beauty as an abstract sculpture).
12th June 2008: 'Grady Booch' said:
i agree...we need well-designed software...but to me, well-designed software is beautiful.
there's an element in my riff re beauty: how do we distinguish good code from bad? in the end, naked code that works, that endures, is well-designed and beautiful. however, how to we teach this?
i often ask academics if they have code reading courses...i've only found two. if you are a civil engineer, you study the works of the masters and critically analyze them. if you are a writer, you study the work of the masters..and so on.
why don't we do this with software? there is no good precedence for "software literary criticism" and by raising the issue of beauty, it hopefully starts making what you mean by "well designed" more manifest.
well-designed software is like pornography (to paraphrase supreme court judge steward): i know it when i see it.
but i think the semantics deserve more than that...
13th June 2008: 'David Wright' said:
Good day Mr. Booch...
The main reason i discuss this whole beauty and/or design question is that it usually accompanies similar issues, such as: is programming an art? and, should the model for programming be craftsmanship or engineering? Its a divisive argument, neither the twain shall meet and all that, and I have pretty much run out of analogies and witty (or not) comments to offer up.
However, you do mention schooling, so I ask this. If i atttend a fine arts school, will prgramming be on the curriculum? Not using software to make art, but actually learning to code?
Now, if I take Engineering, will programming be available on the curriculum? ....
So, I will close (i think) with where I started, "beauty is in the eye of the beholder". Next time I fly, i am not going to care if someone thinks the plane's flight s/w is 'beautiful', but I really hope it works properly...
16th June 2008: 'David Norfolk' said:
Butting in on the conversation between Mr Wright and Mr Booch - I remember that when I started in this game I knew a very fine programmer with a background in languages and Latin.
I have a lot of respect for engineering as a discipline, but building software isn't exactly like building a bridge.
Perhaps we'd get more reliable software if we recruited and trained language specialists - Latinists - to write the stuff, and (perhaps) used engineerrs to specify and implement it...
17th June 2008: 'David Wright' said:
Agreed, developing software is not the same as building a bridge, but engineering as a whole covers many problem domains, and many principles are common across them. Software Enineering is now being included in the curriculums of many Engineering Schools.
...but classical languages as an analogy, or a base to write software on? Interesting... there is also mathematical logic, would that help? I do recall attempts to prove software is correct, using logic I think.
I also recall CASE tools that used a constrained, verifiable pseudologic to define and generate code; they were called Action Diagrams.
17th June 2008: 'David Norfolk' said:
Interesting. Yes, logic is importanty to programming, of course, but a successful "automated business system" is do much more than mere programing logic.
Check out this URL for the effective use of "proof" (but it's a hybrid approach, in which proof is only used where it is appropriate:
http://www.theregister.co.uk/2006/08/14/math_managing_defects/page3.html
Even then, you rely on the spec being corect at the business level.
16th June 2008: 'Grady Booch' said:
when i fly, i don't care if the underlying software is beautiful; if, on the other hand, i am a developer or an organization responsible for the enduring economic value of that software, then i will care about beauty.
immediate functionally is one thing; adaptability, economic viability are others, and there in different measures (such as simplicity, crisp abstractions, a good separation of concern) are relevant.
as a professional, do you desire to write crappy code? if not, then i claim that you yourself have some sense of what beauty is and what it is not.
16th June 2008: 'David Wright' said:
OK, one's opinion on what constitutes 'beauty' cannot be refuted, but a product either performs as designed or it does not. I will take performance over beauty in any product, and enjoy beauty in the arts. Just my opinion...
19th June 2008: 'Jon Collins' said:
While "beauty" might be to strong a word, I think "elegance" might be more appropriate. There is a lot to be said for clean code, whether its in an aeroplane or otherwise - its easier to read, debug etc and its less likely to contain errors. OK, on this last point I have no statistical proof but I can remember the amount of time it took to decode the obfuscations of some programmers before being able to find the bugs that lurked!
Similarly, there is algorithmic and architectural elegance. The best designs are often the simplest, as the patterns guys have so often shown.
Tangentially, a lot of this boils down to the importance of peer review. If you can explain it clearly to a colleague or two, then its probably on the right track even if its not going to win any art foundation awards.
Cheers, Jon
19th June 2008: 'David Norfolk' (Author) said:
Ah yes, peer review is very important to "quality" - and I think that the eXtreme Programming "pair programming" practice institutionalises a form of it....
19th June 2008: 'Jon Collins' said:
Peer review is one of those things thats very easy if its already part of the woodwork, but really tough to institutionalise.
Of course, in my day we practiced pair programming because we couldn't afford a workstation per programmer. Times were tough ;)
19th June 2008: 'David Norfolk' said:
Well, I believe that bench checking code on punched cards before submission to the one compile run for that day (yes, I am that old) really impproves quality - but let's not go back there - please :-)
Of course, what often prevents peer review is the immaturity of many organisations - with management by "creative tension" (fear) and encouragement of a "blame culture"....
The messages above were all contributed by IT-Director.com readers. Whilst we take care to remove any posts deemed inappropriate, we can take no responsibility for these comments. If you would like a comment removed please contact our editorial team.