Slashdot Mirror
HP Lays Off Unix/IA-64 gurus
A reader writes "On Tuesday HP announced that it is closing a lab in NJ. This was an HP-UX development lab, responsible for porting HP-UX to IA64. The lab employed top engineers, including some who have worked in Unix kernels for over 20 years (originally from Bell Labs, Novell, and other companies). " That report came from a soon-to-be former employee.
They've got Compaq. When 64-bit computers become mainstream, they can either go the "Compaq" way and use Windows or TRU-64. Or go IBM's way and stick with Linux. There is really no need to have a third way which is a waste of time porting things in and out. I think R&D should be better spent in improving e-paper or faster scanners/printers.
¦ ©® ±
Makes sense that HP would axe the lesser of the two operating systems.
It's not about products, it's about people. In the R&D business, that's where all the value is. Getting rid of people who are probably in the top 1000 kernel engineers in the world make no sense at all. Why not assign them to merge the best bits of HPUX and Tru64? After all, HP has PA-RISC people, Compaq has Alpha people, but Itanium is a new platform.
This is Fiorina screwing up, again, that's all. I wouldn't be at all surprised if these engineers found a warm welcome waiting for them at Sun or IBM.
HP-UX may be a lesser operating system but it is far more common than Tru64. Tru64 is far faster though.
I especially agree on the "far more common" statement. This is probably just a cost-cutting measure in a market that's not particularly hot right now, but HP should be wary that this might send the wrong message to folks who have committed to HP-UX. Every client I have ever had during my professional career has utilized HP-UX in their network infrastructure to some degree... hopefully they won't get panicky as a result of HP's lack of commitment on IA-64!
"Computer Science is no more about computers than astronomy is about telescopes." - E.W. Dijkstra
At first I was going to agree with you whole heartedly. However, when I thought about it I realized something. In the computer industry it can be difficult to teach old dogs new tricks. Mainframe programmers don't do well in the world of Unix, simularly Apple people don't cross over to Intel well either.
... ).
These guys are incredibily intelligent, but if they don't want to learn something new, it would take a lot of time and money to convince them to do so. I'm sure their salaries were already well above six figures, and it was probably in HP's best interest to let that kind of expense go. They can start fresh with new minds that they can manipulate for a lot less money. It may take a little more time to get them up to speed, but I've got a lot of friends jumping at the chance to play with 'big iron'. They'll work for a lot less money, and get comparable work done in just a little more time. They can also hire four or five new guys for the price of one of the old ones. More man power gives them a larger resource for creativity, more man time, and better 'employee redundancy', which geeks world wide know how great redundancy is.
Business is business, no room for emotion wasted on the trusty old porch dog. Sometimes you need to bring in a new pup (or two, or three, or four
~LoudMusic
No sig for you. YOU GET NO SIG!
Silicon Valley doesn't take SGI seriously any more. Ever since 3D graphics hardware became cheap, SGI has been lost in search of a market niche. They've tried selling servers, creating a Silicon Studio division, making NT workstations, acquiring Cray, getting out of NT workstations, dumping the Silicon Studio division, acquiring Intergraph to get back into NT workstations... Nothing worked. Their basic problem, that their stuff costs 2-3x what comparable stuff costs from others, has yet to be solved.
It will be sad if HP goes that route.
Sounds like you've bought into the major management fallacy of our times - that you can cut one good experienced developer and replace him with 3-4 inexperienced developers for a lower total cost and still get equivalent quality. This is the same fallacy that has led to the wholesale replacement of experienced employees by cheap H1B-visa holders.
The results of the H1B replacement have been extremely poor - almost universally you hear about projects that have gone down the tubes after the transition to H1B. Many times the reason for failure is couched in terms that are not easily linked to the management decision to toss their experienced people, because management is extremely blame adverse. But unless the people are doing the equivalent of "sweat-shop" programming - there is no way that tossing experience in favor of a direct lower cost is going to produce a better product.
At the levels these people people work at, computer science is an art - you make decisions based on prior experience and an instinct based on years of experience discovering what works and what doesn't work. You put a bunch of newbies in there and they will spend the same time taking all the wrong-steps that the experienced people did ten years ago, meanwhile product quality goes out the window and so does time to market. It isn't about teaching old dogs new tricks - the old tricks are fundamental nowadays - just as you don't re-invent the shape of the wheel either.
Plus, if you had read the article, you would see that the people in that lab come from a wide range of backgrounds, they aren't all HPUX crusties - in fact most of them came from Bell Labs just a few years ago. They certainly don't fit the profile of a bunch of old computer geezers who don't know their way around a modern OS or a modern CPU (they were porting to Itanic, some would call that a post-modern CPU - others might call it trash, but that's another story).
When information is power, privacy is freedom.
Of the 120 people, 23 (the best, I assume) were offered jobs in Fort Collins, where a part of the HP-UX work gets done anyway. Also, this is only a fraction of the 1000 or so HP-UX engineers...most of which sit in the Bay Area or Colorado.
The sky is not falling. HP-UX will still be the only non-Linux Unix shipping on Itanium when McKinley rolls around. It looks like Sun and IBM have shelved their ports, for now at least. Don't you think HP gets this?
I've watched fresh college grads who happen to know Java develop a database application from scratch, and it was really sad. No recoverable transactions. No real data structure design. No programming discipline. No documentation. No nothing. I truly feel sorry for the customer who has paid for nothing.
Healthcare article at Kuro5hin
If HP is laying off hackers in NJ, it's not because of their technical value, it's because NJ is Siberia as far as HP is concerned. HP's decision makers are out west. If there is a pile of work to do (and revenue to be made), the suits would rather have it done in their own profit center. And when money is tight, they cut loose the remote location that doesn't have the political clout to defend itself. Having hacked UNIX since the 1970's, I've certainly seen this happen before - I've had it happen to me before.
Its pretty obvious that you're not a senior-level admin...at least not for AIX or HP-UX.
AIX and HP-UX have SO MANY MORE administrative features than Solaris (and let's not even start with Linux) that it's not even funny.
The problem is, there are tons of people just like you, who think that Linux/xBSD rule and don't understand exactly why the big vendors UNIX offerings are truly enterprise-class. So you rip on AIX/HP-UX because you don't know how to effectively manage them. Anybody who is a senior-level admin with either of them can easily be twice as productive with their tasks/chores as on Solaris, or god forbid, Linux.
Don't get me wrong - I love Linux (Slackware!) and the BSDs. But they have their place...and where it is *not* is at the enterprise level.
At least not yet.
sedawkgrep
Is that a salami in my pants or am I just happy to be me?
And part of the real, intangible value being let go is the mentoring of all those fresh college grads. They learn so much more quickly if they work with the grizzled developers (speaking from my experience as a newbie 10 years ago).
Ah, but keeping shareholders happy seems to drive too many management decisions these days...
Peter
Because of that forward product motion, customers could standardize on the HP platform, and buy 3rd party apps and other items that ran under HP/UX (Oracle in particular, since HP/UX is widely used as a base for client/server). With HP/UX 11i as their main server OS, they had some serious scalability and reliability going for them. HP/UX will be supported for the next few years, of course, but once that ends, customers will have the future budgetary choices of sticking with whatever direction Carly takes them in, or abandon HP for a more consistently-managed vendor (i.e. IBM). Bet they pick the latter choice.
I'd have a personalized plate on my car, but "toxic bachelor" won't fit into 7 letters.
I use UNIX, because it is 30+ years old. What other software has aged so gracefully to be:
1) Understandable and predictable from the kernel on up.
2) Immensely useful with a uniform and powerful set of interfaces.
3) Scalable so that my programs work without recompilation from my dinky workstation to that new Sun Fire 15K.
4) Solid as a rock. The only time I have seen Solaris crash was due to a diagnosable and easily fixed mismatch between the video driver and the kernel version.
5) Rewarding. There is always something new to explore in UNIX.
6) Smart. The basic prinicples of UNIX make it a joy to work with (see #7, below).
7) Simple. Yes, UNIX is simple!!!
Healthcare article at Kuro5hin
Here at my job, HP-UX, AIX, Solaris, and WinNT/2K are "tier 1" platforms. All of our products and modules are released and supported for them.
Tru64 simply doesn't have the market share in our segment for us to justify the same level of support. We support it, but in the same way we support Pyramid, Solaris/x86, IRIX, and OpenServer (ie, "barely").
(oh, and this being slashdot, I'll mention that Linux is rapidly gaining acceptance among our customers, and is about halfway to becoming a "tier 1" platform at this point - it will be fully integrated into that list in the next release)
Speaking as one of the HP employees in question, I can say for sure that none of us here would ever work for the Redmond Antichrist. OTOH, we'd also be very wary of working for SuSE or Red Hat, because we don't want to be laid off all over again.
Many of us are considering leaving the profession altogether--we've been kicked around from one company to the next over the last 20 years, and we're sick of it.
The generalization shouldn't be directed at H1Bs, but at the large lowest-bid contracting firms that employed exclusively know-nothing H1Bs. These are the guys that are on yahoo messenger all day talking to somebody back in New Dehli who knows how to make a COM object in VB, or are digging around trying to borrow code that solves a different problem than the one at hand. I've never heard of any firm employing US citizens that were universally unqualified idiots.
These firms were postively infesting the valley during the boom -- although like every other contractor, their business is currently hurting, mainly because things are outsourced direct to the source on the other side of the world.
The smart Indian guys I worked with had nothing but disdane for these guys -- you can buy a bogus programmer degree on the street over there and get yourself a free trip to indentured servitide in Fremont. The guys who did go through some low-end programming school were taught Cobol or Borland Pascal on a 286. Nobody knows enough to check up on them, they trust the scammer outsource firms to do that. Huge projects have gone south because some assumed that there's some process or management structure there, when there isn't -- it's throw 50 monkeys at the problem, and maybe a couple of them will be smart enough to save it. They all bill tho.
The stupid thing about this is that it does cost local jobs for novice programmers (they can't compete with bogus credentials), and furthermore it fucks up the INS system for the people that legitimately should be here on the H1 program.
The latest version of SUSE now includes a new LVM. This LVM uses the same commands and arguments as HP. SUSE has a white paper on the new LVM implementation somewhere on their site.
SUSE also includes the ReiserFS journaling file system. By the way - Linux can store ACLs on most of it's JFS implementations - HP-UX cannot (you can only use ACLs on HFS, not VxFS). Care to explain this brain-damaged design?
Yes, Linux still has problems with enterprise scalability, but not the problems you've mentioned.
p.s. I'm pretty ticked off that RedHat seems to have done nothing with the LVM - not a peep.
I would greatly agree. And in most probability, many of these engineers will mostly be absorbed into other HP labs in other regions; but having the skills, it would not be difficult for them to find good jobs within the area without moving.
There must be some kind of sound reasoning for their decision because HP-UX is still one of their bread and butter divisions.
Return the bells of Balangiga.
if anything , it's a testament to the crappy way big corporations treat loyal and qualified employees
;) All these engineers they're laying off will probably end up with IBM, Red Hat or Sun with an axe to grind. Research lab UNIX (tm) types don't leave the scene to flip burgers at McDonald's. This will come back to bite HP in the ass.
Exactly. This reminds me of all those Digital techies in the Alpha division jumping ship when Compaq took over because their corporate culture sucked and they weren't treated as valuable, talented people. Where did alot of those dudes end up? AMD. And Compaq's blunder has come home to roost against Wintel in the Athlon, with x86-64 as an encore to *really* rub Wintel's face in the dirt.
Now it's HP's turn to step on their dicks....oh I forgot, Carly doesn't have one