Danaher buys Keithley

No. Other than being excellent terminals, PCs were a total waste of time for IBM. That really was their justification in the first place. The original PC was about the same price as a semi-dumb terminal of the time.

Nope. They sell some x-86 servers but the PC business was sold (take it - please!) to Lenovo.

The PC division didn't make money (there were years they did, but overall, no). They're still in business servicing other's PCs, so one supposes they're making money there.

be

Not in Europe and that was one of the problems. Even with roughly half-off for employees the clones were still cheaper.

If they hadn't artificially tried to keep prices very high I think they would have had a chance. Somehow it seemed similar what later happened with a whole genre of chips, the switched capacitor filters. I used them once on a design, where I had to. Tons of other apps later but there I went analog or DSP. The only reason was cost.

be

Huh? IBM's product justification had nothing to do with the competition's prices.

If they hadn't tried to keep their niche they would have lost more money, faster. You may think Microchannel was a loser, but from IBM's perspective, and their customer's, it was exactly what was needed[*]. It did a great job of P-n-P long before such was possible otherwise.

be

PC

They should have paid attention to those though.

they're

Then, who was? I was a consultant most of the time back then so I got around. Never saw much in terms of PS/2 'puters and none of the hardcore board developers was talking MicroChannel.

I was eyeing OS/2 when I was young but there were two obstacles IBM could have done something about. It was expensive, so for someone just starting out self-employed other grass looked greener. Then there were a few weird compatibility issues, some stuff wouldn't work under OS/2 but I needed it to work. Then it all shriveled up, and I wasn't really surprised. Sometimes I wish I could be da big boss at a company just for one year, be given a huge broom, and clean house :-)

they'd be

for

PC

IBM's market was always people with money. ;-) When they got in trouble was when they forgot that.

they're

Oh, "INTERNATIONAL BUSINESSes". ;-) Other than things like CATIA and similar (customers like Boeing) they never did well in the engineering or scientific arenas. Banks wore the same clothes and have more money anyway.

Again, you weren't in their target audience.

I used OS/2 exclusively from DOS7 to Win2K. It worked well enough for me. It kinda got like Linux today; superior, but doesn't work. ;-)

they'd be

for

original PC

they're

They did really well in the mainframe days, IBM had a reputation to make things work. Many of they large competitors often couldn't. But somehow IBM never really thrived in the PC business. Maybe too much key-account thinking on their part?

Some of the companies I consulted for where rather huge. Ok, but if they concentrated on the few even bigger ones then maybe that was a mistake.

It was probably ok when just computing. But folks like me had to hook custom-designed hardware into PCs and there things better be compatible. Beamformers, scan converters, stuff like that.

You had to re-quote his entire post to write that!?

Learn how to snip, dip.

they'd be

for

original PC

all

they're

Oh, there was *definitely* that. In the semiconductor biz there were times where I think they were emulating Maxim, or perhaps going them one better. "No one needs all these damned customers. They cost money!"

If you weren't buying mainframes, or later their consultancy, you weren't big enough.

Custom hardware wasn't a problem, or any more of a problem than it is with Win or Linux today. ATMs ran on it for a couple of decades. Commodity hardware

they'd be

time for

original PC

all

overall,

they're

perspective,

job

And that was the key mistake. Companies of larger and larger sizes no longer needed mainframes. The writing had been on the wall for a long, long time.

Yes, if you make one dedicated product and then not change its design for a decade or so. That was different in our biz, we made sure we were flexible WRT the PC platform. Not so much for pricing reasons but to be able to use the latest and greatest in imaging stations. No way we'd ever even consider a proprietary bus or OS.

Yep, that's the other problem. 50% plus of an ultrasound machine can nowadays be commodity products. Frame grabbers, special graphics cards and so on. That stuff has got to run. Even if a vendor would support some proprietary bus or OS, if they did that with only a few months delay that would have made us miss a major trade show. No way we were going to let that happen.

they'd be

time for

original PC

was

all

it -

overall,

they're

perspective,

job

similar

Except that they're still shipping more and more of them. Dollar-wise, no, but it is still a $10B business (about 10% of the company).

It

The money was saved in the configuration and maintenance costs.

That's why I eventually went to Windows *and* why I haven't gone to Linux. No matter how much the Linuxphools scream about hardware support, it is *not* there.

Joerge thinks you can run a credit card networkwith thousands of ATMs on a couple PCs. ;-)

Well, Windows certainly does have the the largest number of available drivers, but isn't the better question whether or not Linux supports the hardware you're looking to use or not?

If you're building something that has plenty of custom hardware in it anyway -- and hence will need to be writing custom drivers -- it's no longer at all obvious that Windows is the best choice. Heck, look at how many people here complain about the Tek and Agilent scopes that now run Windows!

Even on the application side, I'd argue that many applications (like, oh, say, those used to configure intercom systems) don't typically make use of much proprietary technology within an operating system (e.g., DirectX and COM for PCs, perhaps stuff like named FIFOs in Linux, whatever the Mac OS has [and apparently it has a LOT these days]), etc., and hence if you decide from the start you'd like to make a cross-platform application, it's usually very little additional effort relative to making something that only runs on one platform.

---Joel

With a PC you never know. Once a client had rented an older Rohde&Schwarz portable spectrum analyzer. It had ... ... an IrDA port and some strange proprietary converter came with it. Along with an install CD with, you guessed it, only a Windows installer on there. I'd really been up the creek there with a Linux laptop, couldn't have done my job.

Then, during a meeting at a chip design house lately, my client's engineer wanted to runs his presentation and turned out only USB jacks were in the conf table. No VGA. Dang. So they handed him a CD, "here's the driver, just hit install".

Such gear needs a true RTOS which Windows definitely is not. QNX would be my first choice there. If it really needs an OS. The DSO here is a classic example why this stuff often doesn't really need one. AFAICT it doesn't have any OS yet is fully remote controllable.

But what about drivers for some not so common hardware add-on? That's the most common complaint I heard from Linuxers, "oh, that gizmo would really be great but I can't use it with Linux".

I suppose I figure that unless your contract specifies that you'll be bringing a latop with a specific OS on it with you, you couldn't really be faulted for having a Windowless laptop -- or no laptop whatsoever -- with you, could you? Heck, there are come clients where I expect you wouldn't be allowed to use your own laptop anyway for security reasons, and one would just be provided to you.

Granted, at least with hardware engineers, Windows is probably what you'd

I can't imagine they didn't have a spare laptop around they could have transferred your presentation to if you were sporting, e.g., a Mac?

It is a problem, certainly -- particularly with laptops -- although it's getting better over time. (Unless you figure we'll all still be running Windows 100 years from now, whatever that replacement OS is going to be will surely suffer -- as Linux now does -- with a lack of availability of drivers for some amount of time...).

Keep a Windows-based netbook on hand for those times you need to travel, even if at home your're a hard-core Linux guy, perhaps? :-)

--Joel

Of course it isn't specified but there are some things clients have come to expect from a consultant. That can go quite far. Once an emergency required me to do an overseas trip, right away, ASAP. No time for their travel agent to handle my ticket, considering times zones and all that. They just expected that I plunk down a few thousand for an expensive short notice ticket and then bill that.

That must have been odd, a Sun laptop? Never seen one. Even in the 90's all chip designers I dealt with used PCs. I would have become real suspicious if they hadn't :-)

They did, but he wanted to show all sorts of files and it would have been a pain to piece that scattering together onto a memory stick. We had only a few minutes to get this started, had to head back to an airport.

You could. But it's an extra expense and a laptop is only useful if all the files and apps are kept up. Hard to do if it sits in a corner at home all day long.

Agreed, but I think you might have a somewhat higher set of expectations for that than your clients do.

...or maybe that's why the same people keep extending your contracts? :-)

Exercise that Platium Visa card? Or are you a proud member of the elite who pays $2,500/year (plus a one-time $5,000 joining fee) for the awesomeness of using an American Express Centurion card? :-)

SPARC-based laptops were actually around for quite awhile... even up until a few years ago, it seems:

I don't know the percentages, but my impression is that some significant percentage of Cadence Virtuoso licenses are still for *NIX machines. Certainly that percentage is shrinking, though. Many Cadence tools still run on Solaris as well (e.g,.

The Linux box is supposed to use Wake-On-Lan to power up the Windows box and then synchronize files before you even think about doing it manually.

:-)

Just kidding, but I'm sure *someone* has done something like this. I wouldn't put it past, e.g., Jan Panteltje!

---Joel

I don't think they notice such details much, they just worry about that things get done and problems fixed. Mostly the V.P.Engineering or CEO signs my agreement and then chucks me the keys, "Here, fix it, and report back when it's fixed".

No, I am not a believer in racking up credit card debt. I did have an Amex card for a while but that was under $100/year. Wasn't happy because numerous merchants would not accept it. Fees too high, they said. So I canceled it.

Quote "The unit is really too heavy, and more importantly too hot (especially since you?ll cover the vents), to keep on your lap for extended periods of time" ... Yikes!

And the budget version is $3k, the workhorse sets you back a cool $9k.

Our current chip is done on Mentor. Haven't seen Cadence stuff myself in a long time.

Ok, if someone is retired and has a lot of time, maybe. But I like to keep things simple :-)

Not really. Windows will support any hardware I might buy. Linux, not likely, at lest completely.

Red herring. They shouldn't be running Linux or MAC O/S, either. A scope is an appliance.

Even if cross-platform application programming were free (it most certainly is not), testing is not. Companies support Win only for good reason.

Perhaps not faulted for not bringing a Win laptop, but you *will* be faulted for looking like a damned fool and wasting everyone's time.

By the time the presentation rolls around it's too late. If this is going to be a problem, such arrangements should be made in advance.

I certainly don't see it getting better.

As long as the netbook has all the features of a desktop. ;-)