Slashdot Mirror
The State of R&D At HP, IBM, and Microsoft
jcatcw writes "Computerworld surveys the R&D efforts at HP, IBM and Microsoft ($17 billion annually) and raises the question: Are these companies supporting more long-term basic research, or just the usual short-term, product-oriented work? HP is consolidating its focus on a few 'big bet' projects in five major research areas — information explosion, dynamic cloud services, content transformation, intelligent infrastructure, and sustainability. IBM has four 'high-risk' basic research areas — nanotechnology, cloud computing, integrated systems and chip architecture, and managing business integrity through advanced math and computer science. Many of the 272 research projects named at Microsoft Research's Web site are structured with major product lines like Windows, Office, or Xbox in mind, but many also seem to have no likely application to anything the company sells today."
Dear fucking god, my boss comes in at least once a week and asks me if our flagship app could run on cloud computing. Give me a gun and one bullet please.
These are businesses not institutions. They are in the business of making money via products they have mastered. Not a troll just the truth. However check out some of the better colleges and you'll find some sweet research going on. Then these big companies pay the kid 1m for the rights, patent it and make 100m off of it. Cycle of life :)
Whether research is aimed at solving an immediate problem or a problem conjured up in the mind of a scientist, it is not the technical abilities that make research useful. Any project can come up with a solution, pretty much. Rather, it is the mindset that "we are doing research" that makes the activity so productive.
By opening your eyes to all possibilities and outcomes makes even mistakes useful, and having no penalty (relatively speaking, of course) when those mistakes arise frees researchers to create and build. Instead of creating a tailored solution, they can find various solutions and even branch out into more fruitful areas if the main branch turns cold.
Having overarching themes that you want to pursue, like HP and IBM have, makes it easy for researchers to focus. On the other hand, pure research as they do at MS allows the researchers to go off in any direction that seems fruitful, even if most of the projects end up as dead ends.
...that $35 million line item in Microsoft's budget that reads "EVIL (misc.)"
Evidently, their lab's automated buzzword generating script is being tested.
For all intensive purposes, "whom" is no longer a word. That begs the question, "who cares"?
MS research actually does some pretty cool stuff and, more notably, that cool stuff includes things unrelated to OMG-Must-Keep-Wall-Street-Happy-This-Quarter type projects.
I fully admit that I am speaking from a small dataset; but I am inclined to suspect that good corporate R&D is perhaps the one upside to companies with more market power than is strictly desireable. Think about Bell Labs back in the day. They did loads of basic research, including such minor little projects as transistors and Unix. Compare this to our dear present day carriers, whose primary mode of "innovation" appears to be writing ever more incomprehensible contracts.
This is hardly to say that high concentrations of market power are a good thing, as R&D can be done in startups and universities and the like without all the downsides; but it does seem that high levels of market power do, somewhat, preserve R&D from being sacrificed on the altar of quarterly results.
Didn't researchers at HP recently discover how to build memreistor's? That isn't a 6 month to market product and can profoundly change computing.
Thus negating the article at least for HP. IBM does work like this also it is where the Cell processor idea came from years and years ago.
MSFT does have singularity, and a few other pieces of completely random but cool tech that will break compatibility with their other offerings so they will never see the market but are still done just because.
i thought once I was found, but it was only a dream.
Seriously,
These companies are so large the basic research dies sometime after the researcher gives his presentation to Marketing/Product Management.
Marketing may, actually come up with some good ideas, and send them into some Managerial circle jerk with Sales Management. After a couple of comments, the good ideas die in a perpetual Gordian knot.
The knot presents itself as:
1. No customers for new idea.
2. No budget for some new idea.
3. No "proven" market for new idea.
4. No one is willing to risk their status taking a chance at a new product.
Which leads to a tremendous waste of resources "catching up" to upstarts at a later date.
http://www.maxineudall.com/2010/02/should-economists-be-sued-for-malpractice.html
Parent poster has a narrow view of industry research. I graduated with my PhD in CS about six years ago from a top-20 university and have worked in an industry research lab. The primary output of industry research are patents, papers, and products (either new products or improving products). And the research labs at Microsoft, IBM, HP, and Yahoo are all very good at this. Take a look at the top CS conferences in the fields where these companies have a stake, and you will see that industry research contributes a large share of the paper output (e.g. SOSP, OSDI, SIGMOD, VLDB, WWW, KDD, etc.). Further, these companies are spending lots of money sponsoring a wide breadth of conferences and helping to drive fundamental research at a time when NSF funding is low. These companies should be applauded.
Microsoft R&D do make some stuff that's applicable to current products.
One clear example of that is multi-touch surfaces, that will be supported in Windows 7.
I think that thousands of former IBM employees in North America already know.
So explain why they're hiring?
Yes... they may have reduced their helpless desk staff, but R&D is growing
XML is a known as a key material required to create SMD: Software of Mass Destruction
At one time Dell had a very active R&D. They employed some of the best in the industry. Now they aren't even mentioned. Once the carper baggers got into the company, R&D became a needless expense.
Yes, it is true that large companies spend some money on research. It's not the same kind of spend we used to have though, and that won't change until companies don't have to constantly scramble to post big numbers every quarter.
Since this pressure is going to be there from now on, all research will continue to focus on a product the company can sell in a very short time. (Look at multi-touch surfaces rumored to be included in Windows 7 for example.)
There are two problems that will keep this short term focus going forever. The first is that everyone depends 100% on the stock market for their retirement now -- this wasn't the case in the 60s and 70s when Bell Labs, IBM, etc. were able to invest huge amounts of money in research. The second problem is that because everyone's responsible for their own retirement, they're constantly watching the market, making it impossible for a company to think long term. If you miss your numbers as CEO for more than a quarter or two, you're fired, even if you're doing the right things long-term.
I can't imagine a CEO being able to address a shareholder meeting and explain that research is important. Everyone would start shouting "Shut up and gimme my money!"
"So, what do you think of my program?"
"I wouldn't buy this."
"You wouldn't buy it, because it's designed for engineers."
"Engineers think the same as marketeers."
"If that were true, we would still be in caves wondering if rocks were edible."
"You know, you could keep recipes on this."
"We are Microsoft. You shall be assimilated. Competition is futile."
Sorry I didn't think we were confined to the same company. I was arguing against the idea that the research "dies".
A public example is the Glasgow Haskell Compiler. True, it's not actually Microsoft branded as "MS Visual Haskell" or anything like that, but it is developed largely by Microsoftians (most prominently Simon Peyton Jones at Microsoft Cambridge). We all get a pretty decent and very usable compiler out of it; does it matter so much that it isn't branded as a Microsoft product?
is LEAN. How to get rid of American jobs.
Hmmm ... then what about all this stuff: http://www.research.ibm.com/areas.shtml
Or maybe IBM has secretly invented cell processor AI technology to produce scientific papers ... and "Dr. Who" Cybermen who present them at conferences ...
Note to self: buy more tinfoil, IBM Cybermen are just like totally *everywhere* ...
Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
Sun.
As far as corporate R&D for IT companies, here's my take...
Microsoft: I have no idea what they spend their billions on, and yes, they do spend a bundle on R&D. They just never have much in the way of either (a) actual products to show for it, or (b) wow technologies like HP PARC. Seriously, considering all they spend, they should be HP PARC. They're not.
IBM: Well, they do make the fastest chips on earth at the moment (POWER6), as well as tons of other things, and the best O/S on the planet (z/OS), not to mention another awesome O/S (OS/400), plus...OK, IBM does great things with R&D.
Google: See Microsoft. Seriously - they have a great search engine. Best in the world. Hands down. And also...oh wait, I guess that's it.
HP: HP has pissed away more technology they've bought than anyone else. VMS? Tru64? Alpha? And then there are things they do that are awesome - Tandem Nonstop - but never integrated into the rest of their line. ARGH! They make great printers, I'll give them that...
Sun: Really forward-thinking, great R&D. The Solaris O/S has been Microsoft's crystal ball for years. DTRACE. ZFS. Java. Even just cool implementations like Thumper. Of course, Sun can't seem to make dime one on any of this...they really do seem to have the most clueless management on earth...a pity.
Did I forget anyone?
Advice: on VPS providers