Nathan's laws of software

re: Nathan's laws of software

Ryan — Fri, 17 Jun 2005 20:34:44 GMT

"So in the future, the growth of software is going to become significantly constrained (until some new paradigm comes along)."

I find my self wondering if we're just not going to see a re-emergence (for lack of any better term) of a time when a small number of people could write software that took full advantage of a machine. In my limited experience it seems development has gotten much easier in the last 10 years or so. (The time period I was in college and my professional time.)

With processing capability scaling out rather than up the (currently) there is a limited number of people who can write good code that runs in parallel threads. I'd be fall down shocked to find out there isn’t a bunch of research being done around analyzing today's software to find out how to parallelize common tasks. The next step is pushing the outcome of that research into the dev tools, and then... Of course, that will prove the 4th law.

re: Nathan's laws of software

Chris — Fri, 17 Jun 2005 21:10:19 GMT

Who puts still pictures of themselves talking in a slide show like that? I can see one or two pictures, epsically if they are demonstrating something.

Honestly, I really do *not* want to see that guys face every 3rd or 4th slide. Espically since it is just him talking!

Interesting talk, and interesting post Larry.

re: Nathan's laws of software

Anonymous — Fri, 17 Jun 2005 23:29:13 GMT

There's a touch of hyperbole there. I had a brand new computer in 1997 that was considered a standard desktop class machine (not loaded out) and it had a P2-233, 64MB of RAM, and a 6GB HD. 10GB was an expensive option. Not great shakes, but you could build a respectable MP3 library. (That was pre-WMA of course)

Your point still stands though. My new PocketPC has almost as much capacity and performance as that system did.

re: Nathan's laws of software

LarryOsterman — Fri, 17 Jun 2005 23:43:17 GMT

Chris, I think that there's a version of that that has audio that goes with it - the pictures allow it to make more sense (I think).

I'm not sure what's up with that though.

re: Nathan's laws of software

mgrier — Sat, 18 Jun 2005 07:24:01 GMT

Moore's law is still basically going and will probably continue for a good decade or more. Moore's law is about the number of semiconductor devices that are in a packaged unit (chip).

We're out of ways to make single (x86-compatible) processors faster for the most part but we have 10-20 years of making caches larger and more effective (the storage hierarchy is more important than ever!) and exploiting multiple execution units on a given package (multicore CPUs). Maybe VLIW will finally pay off (or IA64!) and we can get a better IPC without having to go past 3-4ghz for internal CPU clocks. (And yes there's interesting work on asynchronous logic etc. etc. that can pay off; it will be interesting to see if it may be effective enough to ever dislodge the ia32/x64 architectures given the existing investment there...)

What's most interesting is that it looks like the relative costs between the main memory and L1 cache in the storage hierarchy is probably close to an all time high right now; memory busses and available bandwidth are going to be improved during the next two decades also so while I've harboured a belief for quite a while that we need to start treating physical memory the way a database treats the physical disk drive it looks like that will not come to pass.

But that will just highlight the problems between memory and secondary storage; there is certainly interesting work going on there to find a middle ground to the storage hierarchy but unless something really revolutionary happens (remember holographic storage?), rotating media has no replacement for very high volume storage.

I'm sorry we don't have Nathan around any more; I have heard plenty of good and bad things about him but the world around here was a more interesting place with people like him around.

re: Nathan's laws of software

Jerry Pisk — Sat, 18 Jun 2005 10:04:16 GMT

Ryan, I think the reason that so many people can "write" software these days is that the hardware is capable of running hugely inefficient code. For example garbage collectors make writing code simple to large number of people who would not be able to handle their own memory management - at the expense of using (wasting) a lot of hardware resources. There will always be applications that require so much processing power or resources that only a small number of people are capable of writing a code to handle it.

re: Nathan's laws of software

vince — Sat, 18 Jun 2005 19:17:57 GMT

I knew people with massively huge music collections... in 1993.

The wonder of .mod files.

re: Nathan's laws of software

Dan McCarty — Mon, 20 Jun 2005 23:13:32 GMT

"Software is a Gas!"

Not being familiar with the oldster lingo, is that a good thing or a bad thing? ;-)

re: Nathan's laws of software

Dan McCarty — Mon, 20 Jun 2005 23:15:33 GMT

Oh, he actually meant "a gas," like, not a solid! Sorry, scratch that last comment.

What a gas! (Whatever that means...)

re: Nathan's laws of software

LarryOsterman — Mon, 20 Jun 2005 23:15:51 GMT

Dan, don't think old people, think physics.

Nathan's a theoretical physicist, so...

re: MS Office Open XML Formats and OpenDocument XML format

机票 — Wed, 06 Jul 2005 16:47:50 GMT

New Rectangles to the Rescue? (Why the UI, Part 4)

Jensen Harris: An Office User Interface Blog — Tue, 28 Mar 2006 09:22:49 GMT

This is the fourth part in my weekly series of entries in which I outline some of the reasons we decided...

New Rectangles to the Rescue? (Why the UI, Part 4)

Jensen Harris: An Office User Interface Blog — Mon, 03 Apr 2006 17:00:46 GMT

This is the fourth part in my eight-part series of entries in which I outline some of the reasons we...

wordofmansoor.com » Blog Archive » The Laws of Software !!

wordofmansoor.com » Blog Archive » The Laws of Software !! — Tue, 27 Feb 2007 11:02:30 GMT

PingBack from http://mobi.wordofmansoor.com/import/?p=36

Larry Osterman s WebLog Nathan s laws of software | Wood TV Stand

Larry Osterman s WebLog Nathan s laws of software | Wood TV Stand — Mon, 01 Jun 2009 02:42:51 GMT

PingBack from http://woodtvstand.info/story.php?id=1496

Larry Osterman s WebLog Nathan s laws of software | debt consolidator

Larry Osterman s WebLog Nathan s laws of software | debt consolidator — Fri, 19 Jun 2009 17:51:20 GMT

PingBack from http://mydebtconsolidator.info/story.php?id=22943