Well, you're right, it's much cheaper to use a notebook and organizer and calculator.
Thing is, I've got five years of notes in my Visor Edge right now. With my previous organizer I could fit only about two months worth of notes, then I had to remove them and archive them. Need a note from three months ago? Gotta go back to the archive....
Can't search notes in a notebook, either.
Both of these also apply to the calendar. A year's worth of daily calendar pages wouldn't fit, yet I have five years worth of them in the device.
What if you lose the notebook or it gets destroyed? You're SOL. With the device, there's a backup of the data. Get a new device, sit it on the cradle, and a minute later it's like nothing happened (except you're a couple hundred bucks poorer, of course).
These things are real advantages, but there's one more that I just didn't expect: Books. E-books. When I'm taking the bus or the train I always carried a book with me to pass the time. One more thing to carry, you know. The device can hold several full-length novels (I usually carry two, the one I'm reading and one more so there's something to read when I finish the first) plus daily news from a number of sources -- all without adding a bit of bulk.
Nice.
That's all with the Palm, which is ages-old technology at this point. Newer stuff like the Z can fulfill other roles. Mine is doing good duty as an MP3 player on top of its PDA functions. Snap in a CF card full of tunes and listen away, still nothing extra to carry.
The filing date for the patent was 1980, so work
would have to predate that. Luckily Nelson's
hyperlink work started in the mid-1960's. By 1989,
when this was granted, it was widely known. Except,
of course, to the USPTO, and since it's not referenced in the application, presumably to BT either.
We've already had C programming environments that
did this kind of thing. Saber-C was doing it
as early as 1988, and there were at least three
different object code runtime analysis tools
on the market for Sparc by 1994 -- Purify,
TestCenter, and a tool from Sun (whose name
escapes me at the moment). Also by 1994 there
were at least two source-to-source compilers that
would inject runtime error checking into C
code, and a myriad of heap debugging tools. I
never used it, but I understand that SoftICE was
similar for the PC.
It's nice to see this kind of thing come around
again, but it's not like it's new technology.
Still, it'll be a godsend to have it again. Gdb
and Visual C++ suck as debugging tools.
Well, there's certainly some truth to the fact that you get little or poor documentation with most of the open source software. The question I always ask of people who bring this up is, what documentation did you get with the Microsoft software?
I see little or no documentation out of Microsoft for the stuff I buy, either. Nor did I ever get much out of Sun or IBM. When I wanted good documentation I had to go out and buy it -- either from the vendor like in the case of MSDN, or from some book from my bookstore as in the case of X11/UNIX/IBM.
If you're missing documentation for open source products, you should check out your local bookstore. There is actually a remarkable amount of documentation out there if you're willing to spend some money on it. Much of it is crap, of course, same as with the commercial vendors -- but some of it is very very good.
It regularly astounds me that people who were willing to pay thousands of dollars a year for technical information from Microsoft/IBM/Sun/whomever won't spend a dime on the same kind of thing for Linux. Maybe they should. Certainly there are companies that fill this particular niche.
Can someone make money selling docs on Linux? I think they can. They certainly did selling docs on X11, which you might recall was open source too.
>It's not WiFi compliant, which means IT DOESN'T >WORK WITH AIRPORT. It will only work with the >Xircom access point. (Standards? What standards?)
News to me. Works fine with my Lucent (Orinoco)
access point, which is the same 802.11 hardware as
Apple's.
>But wait! It gets better! After biting my tongue >on many bad words and buying the @#($&@!! Xircom >Special Super Secret 802.11 Version access point,
>you know what we found? There's a bug in the >Ethernet->PPP translation layer THAT THROTTLES
>THE CONNECTION TO SERIAL SPEED. Palm acknowledges
>it but has no plans to fix it.
I haven't seen this either, although I'm using the
handspring edge rather than a Palm. Performance
seems pretty good, quite snappy with the Blazer
browser.
They certainly are blocking in my area (Arlington). I tried asking them to unblock it and was stonewalled by the phone support guy. I tried going through their feedback mechanism and got three copies of the same stock message telling me it's against my service agreement to run a web server, when actually it's specifically allowed.
They've pissed off a lot of customers, though, and that's not good for business; I suspect they'll end up going the same route they do with SMB... block it unless specifically requested.
Price comparisons between Windows and Linux are something of an apples-to-oranges comparison. I'll compare it both ways to contrast.
I've bought most of the Red Hat releases since 5.1 -- 5.2, 6.0, 6.1, 7.0, 7.1. I tended to buy the premium boxes (at $70-90), but a couple of times I got the basic package (about $40). All told it comes to about $360, maybe as much as $380.
As a developer I bought MSDN for a few years and got my Windows releases from that. I needed MSDN for my work anyway so it was a pretty good deal. That was $500/year and, since I bought the machine clean of any OS, I didn't pay for the OS twice. That's like $1,500 not including the compilers. I stopped doing that when the MSDN price jumped to $800 (my bill arrived the same week Microsoft was telling the court that they weren't a price gouging monopoly... HAH!).
Since then I've bought Win98 full retail ($190) to upgrade a desktop PC and Win98 and NT were bundled with my laptops (not sure the real cost, but probably on the order of $50 and $150 respectively). I've not upgraded to WinME or Win2K purely because they offer little to no benefit over what I currently run.
So over the course of the last 5 years that I've been using Linux I've paid Microsoft $390 for my operating systems, ignoring for a minute the developer stuff. I've paid pretty much the same for either OS, although I've kept much more current with Linux than Windows. It's not looking so bad for Microsoft.
But there are two problems with that comparison. First, those aren't complete systems -- they're just the OS -- and I'm running fewer Windows systems than Linux systems. To see how much it really cost we have to look at/all/ of the systems, and they must be/usable/ systems.
In terms of installations, I'm running just three Windows-capable systems (one of them dual-boot to Linux). I'm running six Linux systems (again counting the dual-boot). Average price per box is, therefore, about half that of Windows before we even start looking at what it costs to make Windows actually useful.
Now, when I install a Linux system it's pretty much complete. I have gone out and bought Wordperfect ($50, but not used now that Abiword is up to the job) and MTV ($20). $70 on add-on software to do everything I need to do. So far I've only had to buy a single version of these things.
With Windows, however, I bought Office 95 ($200 with a big education discount), Office 98 ($300 with a PC), and virus software ($70) just to get basic functionality. So tack on $570 to the Windows figure, and keep in mind that I was pirating an Office installation for awhile. And, again, I didn't fork out the money for Office 2000 so I'm running a release behind.
Even running without full legal licenses we're now seeing Windows cost more than twice as much per box, with only one system ever being upgraded, as Linux did for six boxes and six upgrades. Effective cost per box if you never upgrade is something like $420 for a WinME installation and $520 for a Win2K installation because Office and antivirus software is so necessary. And that's minimum!
Now, again as a developer, I need compilers. Last time I bought VC++ (5.0 I believe) it cost me $300. So, for a usable developer system, I paid $1,100 for MSDN+Office+VC++. I haven't looked at VC++ prices lately, but last I knew MSDN was $800, so today's prices are $1,400 or more -- and $800 of that comes due again next year.
Compare that to the Linux developer's system. I paid for Red Hat and/nothing else/. Total cost to be up-and-running with a purchased unit is about $40, or $80 if you splurge and get paper docs (which, nowadays, aren't even available with Windows). I splurge, so for my development system I would pay SEVENTEEN TIMES as much. (There's some benefit to VC++ versus what you get with RH, to be sure, but not seventeen times as much benefit no matter how you measure it).
Now, I also run Linux as a server. It runs e-mail and web services (and other things, but those are critical). Were I to do the same on Windows I'd have to buy Win2K Server (what's that, $800 minimum?) to get IIS, plus somebody's mail software (never even tried to cost that out). So for my server box we're looking at a grand or more (a LOT more if I were so stupid as to run Exchange). I opted out of that approach entirely.
So for my wife's box we've got about $350 in Microsoft software (excluding the $70 we spend on ant-virus). For my development box we've got $1,100 in Microsoft software (again no anti-virus). For my laptop we've got $50 in Microsoft software (running Win98 w/o Office or anti-virus since I only use it to play DVDs). That's $1,500 in Microsoft software... not including upgrades, and not including server functionality... to get our desktop and laptops doing what we need. And I'm years out of date with VC++ (possible only because I do most development in Java nowadays).
Compare that the $80 I just spent on RH 7.1. I don't need WordPerfect anymore, but I like MTV, so that gives me a total system cost of $110, and that covers three laptops and three servers. And, if I wanted, I could/legally/ reduce that price to $20.
Given that I have not yet spent as much money on Red Hat software, over five years, as I paid for a single year of MSDN I would have to say that yea, Microsoft software is more expensive... so long as you're comparing fully useful systems. If all you're doing is comparing the base operating system then Windows comes out pretty much even, although I tend to upgrade only about a third as often.
The numbers I've paid to Microsoft start to get really scary when I add them all up. Conversely, the numbers I've paid Red Hat were so low that it's hard to see how they will ever make much money.
If I were Microsoft I'd give financial incentives to the PC vendors to not include Java. They've used Windows pricing as a lever in many cases before (look at what they did to IBM) and there's nothing stopping them from doing it again.
While they are restricted against denying the manufacters the right to change the desktop, they aren't restricted in how they can price things, nor what they can do with marketing money. And money is a huge bargaining chip.
As a result I would be surprised if any of the large PC manufacturers ended up bundling the JRE.
I also don't think it much matters, Java is terrible in the browser (this from someone who has been writing Java almost exclusively since 1995). If I'm writing applications rather than applets I'll just ship the JRE with the installer.
You have to keep in mind that the more nodes you add in a wireless LAN (or any other shared medium LAN), the more contention you get. For doing
a cluster, the practical limit would likely be in the neighborhood of 10 machines...
This is only true if you have only a single access point (and a cheap one at that). You can add bandwidth by adding multiple cards to an access point (assuming it supports it; the WaveLAN APII hardware supports up to two) and/or by adding more access points (increasing AP density).
Our corporate net supports hundreds of wireless systems with very acceptable performance.
Certainly, I ran it that way for most of last year.
But, frankly, using a Linux box as the gateway system for these things is more difficult to administer and not much cheaper (if at all) than using dedicated hardware. When a wireless bridge and NAT router (such as the Lucent Orinoco RG-1000 or the Apple Airport hub) costs under $300 they're a great deal.
My wireless system, which started out using WLAN on a Linux system that was also doing IP masquerading for my cable modem, eventually migrated to the Orinoco AccessPoint II for range and ease-of-administration improvements. (I also retired the box for IP masquerading when hardware NAT boxes dropped well under $200; now it's just a server.)
Let me tell you, wireless rocks for basic net usage. Performance, while well below 100baseT, is more than good enough for two laptops running simultaneous full-bandwidth RealVideo feeds. For home use it's hard to complain about that.
My company rolled it out to all laptop users last year, too, so I've seen how it scales -- and the answer is "pretty darn well." Bulk downloads are fairly slow, but day-to-day use is indistinguishable from the wired connection.
What the hell are you talking about? Solaris and the SPARC architecture are the two biggest
assets Sun has and serving Sun really well. Just take a look at Sun's stock.
Assets or not, stock price is not a reliable indicator of strength. One bad quarter can demolish the value of a company; look at how AAPL did.
As for whether or not SPARC and Solaris are assets, you're looking at it in hindsight. They certainly have been assets, but consider:
SPARC systems and Solaris used to dominate workstation and small server sales. Today they are almost nonexistant in those roles, having been replaced by Intel systems running a variety of operating systems (primarily NT, but increasingly Linux).
SPARC systems and Solaris have been making more and more money lately, but they've been doing it selling larger and larger systems. They're selling because, at the moment, it's cheaper to buy a bigger system than to figure out how to distribute smaller ones. This is particularly the case for databases.
At some point, however, these systems will no longer scale. When that happens it is no longer to the advantage of the buyer to buy a couple of big systems at a premium when a bunch of smaller ones can be had at commodity prices. When that happens the bottom will fall out of Sun's market.
This is exactly what befell IBM with the onslaught of minicomputers in the 1970s, although IBM was diverse enough that it took a long time before it affected their stock. Things went much worse for DEC; they collapsed practically overnight as small, cheap systems from the workstation vendors (particularly Sun) demolished their sales.
Sun cannot survive indefinitely by selling larger and larger machines, and they will come under increasing price pressure on lower-end machines as they get faster and larger. Either Sun gets a lot more efficient or they're going to take a fall like DEC.
Hmm, my Cobalt Qube 2 doesn't even support NFS, making me wonder how useful NFS is as a performance differentiator for this kind of box.
Rather, consider web page serving and e-mail statistics, which are prime features for the Cobalt systems. Solaris/Intel gets pretty well creamed by Linux doing those things -- and Linux does it with a lot less hardware.
But generally I don't think that the observed performance will be that different with either system. I think the big question is whether or not most Cobalt users use the systems out-of-the-box. If they do, then swapping Solaris for Linux isn't going to make much of a difference. But if they're like me and they took their Cobalt box and added a bunch of stuff to it, well, Solaris is a lot more painful for that kind of thing (particularly if Sun continues its practice of not shipping a compiler).
But hey, it's not really about what's good for the user anyway. It's about branding. Linux ain't Sun's brand.
'm a huge Linux fan and all, but let's face it, Solaris is still a fundementally better operating system. I mean, there is no way that
Linux will ever run as efficiently on an UltraSparc chip as Solaris. Much like Solaris x86 on Intel isn't that great.
Just out of curiousity, why couldn't Linux be as efficient as Solaris on UltraSparc? Frankly, having used both systems extensively, I find that Solaris is not especially efficient. You don't notice that so much nowadays because you're buying a ton of horsepower when you buy one of today's SPARC systems, but it's true. Just compare BSD to Solaris on one of their workstation systems -- BSD screams in comparison.
You mention Solaris' as being relatively poor on the x86. How different do you think Solaris/Intel really is from Solaris/SPARC? They're almost entirely the same code, and Sun actually did a pretty good job with their Intel port. But it's nowhere near as efficient as BSD or Linux so it feels like a dog.
I have no doubt that Sun can pare down Solaris to be a good match for the Cobalt hardware, but I question whether this is a good use of resources and whether it's going to be good for customers. The Cobalt systems are beautiful -- incredibly easy to set up, efficient, and extensible using readily available, and easily installed, Linux software. Switch that to Solaris and you have an instant porting job for additional software, and that assumes that Sun does an about-face and starts bundling compilers back in with the OS (fat chance). As both a Cobalt and Sun customer I'm dead set against this kind of change.
Anyway, having said all of that, the Cobalt purchase makes a hell of a lot of sense. It is, in fact, the first major move I've seen them make in years that looks to be ahead of the market rather than behind. They've been making good money selling bigger and bigger servers, but they've been losing an increasing chunk of the lower end of the market. The problem with that is that Moore's Law guarantees that the jobs that you need a monster machine for today will be doable on commodity stuff tomorrow. The big machine user base will eventually peter out (just as it did for Cray and the other supercomputer manufacturers).
Turnkey server systems are going to be a big thing, I think, in a market that Sun has otherwise lost.
Given that a lot happened that year, it's not really clear that the IBM is the first. But Atlas, same year, was certainly the most advanced, pioneering quite a lot of the features that wouldn't appear elsewhere for a decade or more.
If nothing else, you want to stay away from Bell Atlantic as your DSL provider.
Part of the problem is that they take every opportunity to drag their feet when a DSL reseller is involved, but don't think that you'll get a lot better service if you're dealing with them directly. Nearly everyone that I know who has used BA as a DSL provider considers their attitude towards their DSL customers as a kind of uninformed apathy.
If nothing else they're doing a really terrific job of illustrating the very worst aspects of a monopoly.
but Linux, a bunch of 14 year olds, leading the world makes me want to call my mommy.
Maybe you should converse with some of the people doing this work. There aren't too many "uneducated... hackers" working on it so far as I can tell. Virtually every Linux hacker I know has formal education and works with Linux as a hobby. (And since it's a hobby they're more inclined to do it right than to ship something by June 1st, if you get my drift.)
In the end, though, who cares who wrote it so long as it gets the job done? I mean, you're assuming that the guys who put NT together were well educated, that any education they had was actually useful, and also that even if both are true they'll do a good job. That's a lot of very questionable assumptions, particularly about employees who are known to have built MFC and that brain-dead FIFO page replacement algorithm used in NT. (Ok, the latter made a certain amount of sense on the VAX. But couldn't they have done something better now that they have page reference bits?)
Now, bunches of NT people are sitting there thinking that I'm one of those Linux loonies, and there's certainly some bias on my part towards UNIX, but I was writing articles for UNIX people about the viability of NT back when NT was considered a joke by pretty much everyone in the business. NT is a damn fine workstation OS, particularly in a world where the majority of software is written for Windows, and I still use and recommend it in a lot of situations. Back in 1994 I figured that it'd decimate the UNIX workstation market -- and it did.
So NT isn't poison to me, but I have serious reservations about it as a server, most particularly because its stability isn't so hot, but also because I haven't been all that fond of how much I've had to spend to get extra software to do things that UNIX has done out of the box for years.
Yea, yea, I hear you saying "Win2K fixes the stability problem." So Microsoft claims, and maybe it's even true, but given what it's doing its hardware requirements are a little out of hand and the cost... well, I think Microsoft is taking a lot of people for a ride. For a lot of server functions you can buy the OS and hardware for less than Win2K alone.
Then again, I'm educated enough to realize that I have a choice in the matter, and perhaps that's the real threat of Linux. It's not the 14 year olds, it's the guys who are smart enough to be comfortable with not using Windows. There are a lot of those guys, both old-timers and recent graduates, in IT shops and software development houses. Those are the guys who made Linux grow so much in the server space last year.
Maybe Linux really is made largely by 14 year olds and I've just not run into them. So what if it is? It's cheap, it's stable, and it has a hell of a lot of functionality. It's not always the best choice for the job, but you're stupid not to at least consider it.
Similarly, sometimes you have to bend over backwards to get Linux to do the job. This is particularly the case for a lot of specialized applications. So look around you and see what works best for what you have to do.
I suspect that for a lot of people that'll be a mix of OSs. It certainly is for me.
I have to agree with the author that X needs a serious rendering model refresh. I'm not all that willing to concede that X's rendering model is largely the result of hardware limitations at the time of its conception, though; a lot of its problems were really well known and correctable right from the start. There was, shall we say, resistance to those kinds of changes by some of the people involved.
I'd like to add one additional request in the name of bandwidth reduction. Display lists. One of the more interesting things about NeWS (Sun's post-SunView windowing system based on PostScript) was its ability to be programmed. You could create a routine that could render a whole button with bevels and shading and whatnot as a single function call whereas in X it will typically take six or more protocol requests to get the same job done.
While full programmability ala NeWS is rife with potential problems, given just the ability to record a sequence of parameterized protocol requests and request playback would drastically reduce the bandwidth requirements and boost the performance of a lot of applications while having very little impact on memory use. I believe display lists were considered and discarded originally due to the requirement that X11 run in a terminal with very little memory (too little to hold much in the way of display lists). I've never appreciated that since the client-side fallback technique was always a viable option, but given the opportunity to extend the protocol I think this would be something to consider. jim frost
The main reason to use Java is NOT portability
on
IBM JDK 1.3 For Linux
·
· Score: 1
Sun marketing be damned, the main reason to use Java is not portability, it's development time. Versus C or C++ my productivity at least tripled and for certain kinds of applications (network servers in particular) it's way, way better than that. And that's been pretty much the same story for everyone else I know using it, too. Furthermore the bug counts are way, way down and the bugs tend to be irritating rather than debilitating (ie throw an exception and keep running rather than dumping core).
Swing and AWT could use a lot more maturity but if you don't need a GUI, or you only need a basic GUI, Java is awfully nice for building applications.
More speed I can definitely use but realistically speaking we're building some damn fast servers with Java, stuff that runs rings around the C++ competition, even using the Sun 1.1 JDKs. When you're building the code three times as fast you have a lot more time for optimization!
Reading through a lot of these comments makes me believe that my own experience with computers in the classroom is not unique.
They didn't help. Mostly, they were ignored.
If we're going to use computers pervasively in the classroom we need to use them as information resources -- modern textbooks -- not full-blown computational systems. It's just not cost-effective to give general-purpose systems to everyone.
When I was in high school in the early 80s the history books were printed during the Korean War. I didn't learn about Watergate until I was a senior (1984) and it's not like I was taking the basketweaving classes. Twenty year old textbooks were the norm! They couldn't afford to replace the textbooks more often than that.
Given that this is the case, any electronics in the classroom have to have economic benefit or it will simply not happen on a broad scale.
It's not real hard for me to see some possible economic incentive for giving everyone a $50 electronic book and having them download their textbooks. With publisher subsidies (and there are some good economic reasons why they might be willing to subsidize) these devices could break the $20 mark at the consumer level within five years. There are already workable readers selling for under $100 today so I think this is very achievable.
Since most of the cost of a textbook is in printing and shipping the end cost of outfitting a student with all the necessary texts would drop by an order of magnitude, enough that they could have up-to-date textbooks and still save the school money. And the publishers would have better margins, too.
Now, I can believe that there would be some value to the flexibility of a laptop or other general-purpose unit, but let's be realistic here. Public schools are driven primarily by cost and secondarily by benefit. If information technologies can reduce costs as well as provide benefit they'll go with them; if not, they'll keep using ancient texts.
i got DSL, and with my 128Kbit upload, im gonna be spreadin all these decrypted DVDs all over the internet
Given the broadly available connection technology you are correct, DVDs are not really instantly distributable. But he has a point: it won't stay that way forever. It probably won't stay that way for more than ten years.
What they're doing is trying to forestall the problem in the future. This is forward-thinking and exactly what they need to do to protect their businesses. It's really hard to have a beef with that.
Where they're missing the boat is that they can't stop the technology from being published no matter what they do. There are just too many jurisdictions and everyone with a connection is a publisher. It's a battle they lost the day DeCSS was published. Maybe if it had happened earlier in the life of DVD players they could do something about it, but now they have the choice of wimpering about it or making everyone buy new DVD players. The latter won't happen now that DVDs have exceeded critical mass.
Rather, they're going to need to come up with approaches to track down the offenders, just like they do for existing analog piracy. Nothing new here.
Now, he's right that digital distribution is going to make it a lot easier for the thieves to distribute their stuff. But -- and this is really important -- it's going to make it easier for the content producers to distribute it too. The cost of distribution is falling through the floor, and economies of scale will still work in their favor just as they did with videotape.
They're still going to make money in a world where instantaneous distribution is possible. They just have to find ways to make it worth the consumer's while to buy it rather than steal it, and by and large that means they have to keep it affordable.
If you ever look at the way Sun competes against NT, they push their "one big-ass server" approach with the E10000, and says that NT/Intel has no solution in that area. Microsoft is trying to say that the many-server approach that NT backends take is more reliable than the "one big-ass server" that Sun wants to sell you.
I don't agree with that approach either, but there's nothing stopping you from buying a cluster of smaller Suns (we recommend this approach) and since the base availability is much better (as is the manageability) it's a lot easier to keep running and you can use a much smaller cluster.
Having said that I prefer lots of smaller machines myself, but not running NT. You have to do too many things at the NT console, and reinstalling NT is really painful, and they just don't stay up as long as they should. A cluster of BSD machines is a rather cost-effective approach to this problem (isn't that what Yahoo and HotMail do?). At least you know that when one goes down it'll be because of a hardware failure, and you know you can reload the thing from scratch in about a half hour.
I think I would like to see some of the details about their previous Sun infrastructure and its downtime issues. I'd also like to know what the downtime numbers have been for the NT cluster, and how much data they really have (looks like they only have about 2 months of data).
In my experience (and what my company does is high-volume web applications) NT servers are really unreliable relative to Sun. It's not even a contest -- it's an order of magnitude difference. Furthermore, there are real limits to the size and power of a single NT machine relative to Solaris.
The article talks about serving 8.9M pages in a day as peak load. They don't tell you how many of these machines they're running, but we do better than that on one 4-CPU Sun 5000.
I find their harping on eBay pretty funny. Yea, eBay has had some really bad outages on Sun hardware, but that's not so much the fault of Sun as it's the fault of a really bad application architecture on the part of eBay. You're stupid if you think that you can depend on a single piece of hardware or software (no matter how good it is) never failing. It's something of a credit to Sun that eBay is as reliable as it is, and a credit to Sun that they were able to scale their architecture to the degree they have.
Our stuff runs dandy on both NT and Solaris, but customer feedback indicates that Solaris has vastly better reliability.
...what I observed, is when the "One big box" approach fails, and the software, you are talking about, is rewritten to be destributed - that other box one buys often is still a Sun.
Certainly, especially on the first iteration. (After all, you already have that monster box.) But it isn't always a Sun, and that means eroding market share. And even when it is a cluster of Suns you have to buy quite a lot of them before Sun makes as much money as when you buy a single Starfire.
Smaller boxes are bad for Sun. They're bad because you can buy them from some other vendor, and they're bad because they have smaller profit margins. And, like it or not, smaller and cheaper is the natural progression in this business.
Slashdot Mirror
Well, you're right, it's much cheaper to use a notebook and organizer and calculator.
Thing is, I've got five years of notes in my Visor Edge right now. With my previous organizer I could fit only about two months worth of notes, then I had to remove them and archive them. Need a note from three months ago? Gotta go back to the archive....
Can't search notes in a notebook, either.
Both of these also apply to the calendar. A year's worth of daily calendar pages wouldn't fit, yet I have five years worth of them in the device.
What if you lose the notebook or it gets destroyed? You're SOL. With the device, there's a backup of the data. Get a new device, sit it on the cradle, and a minute later it's like nothing happened (except you're a couple hundred bucks poorer, of course).
These things are real advantages, but there's one more that I just didn't expect: Books. E-books. When I'm taking the bus or the train I always carried a book with me to pass the time. One more thing to carry, you know. The device can hold several full-length novels (I usually carry two, the one I'm reading and one more so there's something to read when I finish the first) plus daily news from a number of sources -- all without adding a bit of bulk.
Nice.
That's all with the Palm, which is ages-old technology at this point. Newer stuff like the Z can fulfill other roles. Mine is doing good duty as an MP3 player on top of its PDA functions. Snap in a CF card full of tunes and listen away, still nothing extra to carry.
It's nice to see this kind of thing come around again, but it's not like it's new technology. Still, it'll be a godsend to have it again. Gdb and Visual C++ suck as debugging tools.
I see little or no documentation out of Microsoft for the stuff I buy, either. Nor did I ever get much out of Sun or IBM. When I wanted good documentation I had to go out and buy it -- either from the vendor like in the case of MSDN, or from some book from my bookstore as in the case of X11/UNIX/IBM.
If you're missing documentation for open source products, you should check out your local bookstore. There is actually a remarkable amount of documentation out there if you're willing to spend some money on it. Much of it is crap, of course, same as with the commercial vendors -- but some of it is very very good.
It regularly astounds me that people who were willing to pay thousands of dollars a year for technical information from Microsoft/IBM/Sun/whomever won't spend a dime on the same kind of thing for Linux. Maybe they should. Certainly there are companies that fill this particular niche.
Can someone make money selling docs on Linux? I think they can. They certainly did selling docs on X11, which you might recall was open source too.
>It's not WiFi compliant, which means IT DOESN'T >WORK WITH AIRPORT. It will only work with the >Xircom access point. (Standards? What standards?)
News to me. Works fine with my Lucent (Orinoco)
access point, which is the same 802.11 hardware as
Apple's.
>But wait! It gets better! After biting my tongue >on many bad words and buying the @#($&@!! Xircom >Special Super Secret 802.11 Version access point,
>you know what we found? There's a bug in the >Ethernet->PPP translation layer THAT THROTTLES
>THE CONNECTION TO SERIAL SPEED. Palm acknowledges
>it but has no plans to fix it.
I haven't seen this either, although I'm using the
handspring edge rather than a Palm. Performance
seems pretty good, quite snappy with the Blazer
browser.
YMMV, but the unit works pretty well for me.
They certainly are blocking in my area (Arlington). I tried asking them to unblock it and was stonewalled by the phone support guy. I tried going through their feedback mechanism and got three copies of the same stock message telling me it's against my service agreement to run a web server, when actually it's specifically allowed.
... block it unless specifically requested.
They've pissed off a lot of customers, though, and that's not good for business; I suspect they'll end up going the same route they do with SMB
If not, bye-bye AT&T.
Price comparisons between Windows and Linux are something of an apples-to-oranges comparison. I'll compare it both ways to contrast.
... HAH!).
/all/ of the systems, and they must be /usable/ systems.
/nothing else/. Total cost to be up-and-running with a purchased unit is about $40, or $80 if you splurge and get paper docs (which, nowadays, aren't even available with Windows). I splurge, so for my development system I would pay SEVENTEEN TIMES as much. (There's some benefit to VC++ versus what you get with RH, to be sure, but not seventeen times as much benefit no matter how you measure it).
... not including upgrades, and not including server functionality ... to get our desktop and laptops doing what we need. And I'm years out of date with VC++ (possible only because I do most development in Java nowadays).
/legally/ reduce that price to $20.
... so long as you're comparing fully useful systems. If all you're doing is comparing the base operating system then Windows comes out pretty much even, although I tend to upgrade only about a third as often.
I've bought most of the Red Hat releases since 5.1 -- 5.2, 6.0, 6.1, 7.0, 7.1. I tended to buy the premium boxes (at $70-90), but a couple of times I got the basic package (about $40). All told it comes to about $360, maybe as much as $380.
As a developer I bought MSDN for a few years and got my Windows releases from that. I needed MSDN for my work anyway so it was a pretty good deal. That was $500/year and, since I bought the machine clean of any OS, I didn't pay for the OS twice. That's like $1,500 not including the compilers. I stopped doing that when the MSDN price jumped to $800 (my bill arrived the same week Microsoft was telling the court that they weren't a price gouging monopoly
Since then I've bought Win98 full retail ($190) to upgrade a desktop PC and Win98 and NT were bundled with my laptops (not sure the real cost, but probably on the order of $50 and $150 respectively). I've not upgraded to WinME or Win2K purely because they offer little to no benefit over what I currently run.
So over the course of the last 5 years that I've been using Linux I've paid Microsoft $390 for my operating systems, ignoring for a minute the developer stuff. I've paid pretty much the same for either OS, although I've kept much more current with Linux than Windows. It's not looking so bad for Microsoft.
But there are two problems with that comparison. First, those aren't complete systems -- they're just the OS -- and I'm running fewer Windows systems than Linux systems. To see how much it really cost we have to look at
In terms of installations, I'm running just three Windows-capable systems (one of them dual-boot to Linux). I'm running six Linux systems (again counting the dual-boot). Average price per box is, therefore, about half that of Windows before we even start looking at what it costs to make Windows actually useful.
Now, when I install a Linux system it's pretty much complete. I have gone out and bought Wordperfect ($50, but not used now that Abiword is up to the job) and MTV ($20). $70 on add-on software to do everything I need to do. So far I've only had to buy a single version of these things.
With Windows, however, I bought Office 95 ($200 with a big education discount), Office 98 ($300 with a PC), and virus software ($70) just to get basic functionality. So tack on $570 to the Windows figure, and keep in mind that I was pirating an Office installation for awhile. And, again, I didn't fork out the money for Office 2000 so I'm running a release behind.
Even running without full legal licenses we're now seeing Windows cost more than twice as much per box, with only one system ever being upgraded, as Linux did for six boxes and six upgrades. Effective cost per box if you never upgrade is something like $420 for a WinME installation and $520 for a Win2K installation because Office and antivirus software is so necessary. And that's minimum!
Now, again as a developer, I need compilers. Last time I bought VC++ (5.0 I believe) it cost me $300. So, for a usable developer system, I paid $1,100 for MSDN+Office+VC++. I haven't looked at VC++ prices lately, but last I knew MSDN was $800, so today's prices are $1,400 or more -- and $800 of that comes due again next year.
Compare that to the Linux developer's system. I paid for Red Hat and
Now, I also run Linux as a server. It runs e-mail and web services (and other things, but those are critical). Were I to do the same on Windows I'd have to buy Win2K Server (what's that, $800 minimum?) to get IIS, plus somebody's mail software (never even tried to cost that out). So for my server box we're looking at a grand or more (a LOT more if I were so stupid as to run Exchange). I opted out of that approach entirely.
So for my wife's box we've got about $350 in Microsoft software (excluding the $70 we spend on ant-virus). For my development box we've got $1,100 in Microsoft software (again no anti-virus). For my laptop we've got $50 in Microsoft software (running Win98 w/o Office or anti-virus since I only use it to play DVDs). That's $1,500 in Microsoft software
Compare that the $80 I just spent on RH 7.1. I don't need WordPerfect anymore, but I like MTV, so that gives me a total system cost of $110, and that covers three laptops and three servers. And, if I wanted, I could
Given that I have not yet spent as much money on Red Hat software, over five years, as I paid for a single year of MSDN I would have to say that yea, Microsoft software is more expensive
The numbers I've paid to Microsoft start to get really scary when I add them all up. Conversely, the numbers I've paid Red Hat were so low that it's hard to see how they will ever make much money.
jim frost
If I were Microsoft I'd give financial incentives to the PC vendors to not include Java. They've used Windows pricing as a lever in many cases before (look at what they did to IBM) and there's nothing stopping them from doing it again.
While they are restricted against denying the manufacters the right to change the desktop, they aren't restricted in how they can price things, nor what they can do with marketing money. And money is a huge bargaining chip.
As a result I would be surprised if any of the large PC manufacturers ended up bundling the JRE.
I also don't think it much matters, Java is terrible in the browser (this from someone who has been writing Java almost exclusively since 1995). If I'm writing applications rather than applets I'll just ship the JRE with the installer.
jim frost
You have to keep in mind that the more nodes you add in a wireless LAN (or any other shared medium LAN), the more contention you get. For doing
a cluster, the practical limit would likely be in the neighborhood of 10 machines...
This is only true if you have only a single access point (and a cheap one at that). You can add bandwidth by adding multiple cards to an access point (assuming it supports it; the WaveLAN APII hardware supports up to two) and/or by adding more access points (increasing AP density).
Our corporate net supports hundreds of wireless systems with very acceptable performance.
jim frost
Certainly, I ran it that way for most of last year.
But, frankly, using a Linux box as the gateway system for these things is more difficult to administer and not much cheaper (if at all) than using dedicated hardware. When a wireless bridge and NAT router (such as the Lucent Orinoco RG-1000 or the Apple Airport hub) costs under $300 they're a great deal.
My wireless system, which started out using WLAN on a Linux system that was also doing IP masquerading for my cable modem, eventually migrated to the Orinoco AccessPoint II for range and ease-of-administration improvements. (I also retired the box for IP masquerading when hardware NAT boxes dropped well under $200; now it's just a server.)
Let me tell you, wireless rocks for basic net usage. Performance, while well below 100baseT, is more than good enough for two laptops running simultaneous full-bandwidth RealVideo feeds. For home use it's hard to complain about that.
My company rolled it out to all laptop users last year, too, so I've seen how it scales -- and the answer is "pretty darn well." Bulk downloads are fairly slow, but day-to-day use is indistinguishable from the wired connection.
Wireless will change your usage pattern.
jim frost
Assets or not, stock price is not a reliable indicator of strength. One bad quarter can demolish the value of a company; look at how AAPL did.
As for whether or not SPARC and Solaris are assets, you're looking at it in hindsight. They certainly have been assets, but consider:
- SPARC systems and Solaris used to dominate workstation and small server sales. Today they are almost nonexistant in those roles, having been replaced by Intel systems running a variety of operating systems (primarily NT, but increasingly Linux).
- SPARC systems and Solaris have been making more and more money lately, but they've been doing it selling larger and larger systems. They're selling because, at the moment, it's cheaper to buy a bigger system than to figure out how to distribute smaller ones. This is particularly the case for databases.
Sun cannot survive indefinitely by selling larger and larger machines, and they will come under increasing price pressure on lower-end machines as they get faster and larger. Either Sun gets a lot more efficient or they're going to take a fall like DEC.At some point, however, these systems will no longer scale. When that happens it is no longer to the advantage of the buyer to buy a couple of big systems at a premium when a bunch of smaller ones can be had at commodity prices. When that happens the bottom will fall out of Sun's market.
This is exactly what befell IBM with the onslaught of minicomputers in the 1970s, although IBM was diverse enough that it took a long time before it affected their stock. Things went much worse for DEC; they collapsed practically overnight as small, cheap systems from the workstation vendors (particularly Sun) demolished their sales.
jim frost
Rather, consider web page serving and e-mail statistics, which are prime features for the Cobalt systems. Solaris/Intel gets pretty well creamed by Linux doing those things -- and Linux does it with a lot less hardware.
But generally I don't think that the observed performance will be that different with either system. I think the big question is whether or not most Cobalt users use the systems out-of-the-box. If they do, then swapping Solaris for Linux isn't going to make much of a difference. But if they're like me and they took their Cobalt box and added a bunch of stuff to it, well, Solaris is a lot more painful for that kind of thing (particularly if Sun continues its practice of not shipping a compiler).
But hey, it's not really about what's good for the user anyway. It's about branding. Linux ain't Sun's brand.
jim frost
Just out of curiousity, why couldn't Linux be as efficient as Solaris on UltraSparc? Frankly, having used both systems extensively, I find that Solaris is not especially efficient. You don't notice that so much nowadays because you're buying a ton of horsepower when you buy one of today's SPARC systems, but it's true. Just compare BSD to Solaris on one of their workstation systems -- BSD screams in comparison.
You mention Solaris' as being relatively poor on the x86. How different do you think Solaris/Intel really is from Solaris/SPARC? They're almost entirely the same code, and Sun actually did a pretty good job with their Intel port. But it's nowhere near as efficient as BSD or Linux so it feels like a dog.
I have no doubt that Sun can pare down Solaris to be a good match for the Cobalt hardware, but I question whether this is a good use of resources and whether it's going to be good for customers. The Cobalt systems are beautiful -- incredibly easy to set up, efficient, and extensible using readily available, and easily installed, Linux software. Switch that to Solaris and you have an instant porting job for additional software, and that assumes that Sun does an about-face and starts bundling compilers back in with the OS (fat chance). As both a Cobalt and Sun customer I'm dead set against this kind of change.
Anyway, having said all of that, the Cobalt purchase makes a hell of a lot of sense. It is, in fact, the first major move I've seen them make in years that looks to be ahead of the market rather than behind. They've been making good money selling bigger and bigger servers, but they've been losing an increasing chunk of the lower end of the market. The problem with that is that Moore's Law guarantees that the jobs that you need a monster machine for today will be doable on commodity stuff tomorrow. The big machine user base will eventually peter out (just as it did for Cray and the other supercomputer manufacturers).
Turnkey server systems are going to be a big thing, I think, in a market that Sun has otherwise lost.
jim frost
I agree with you that modem-only connectivity is a loser, but I don't agree that Ethernet is necessarily the right way to go.
You have this ultra-portable pad thing, and you want to wire it to the net? That's insane. These things cry for wireless. And it's not even an option.
Count me out of this one. (And that's too bad, 'cause the first company that gets it right will have me stampeding for the checkout counter.)
jim frost
...anybody have a copy?
jim frost
Given that a lot happened that year, it's not really clear that the IBM is the first. But Atlas, same year, was certainly the most advanced, pioneering quite a lot of the features that wouldn't appear elsewhere for a decade or more.
jim frost
If nothing else, you want to stay away from Bell Atlantic as your DSL provider.
Part of the problem is that they take every opportunity to drag their feet when a DSL reseller is involved, but don't think that you'll get a lot better service if you're dealing with them directly. Nearly everyone that I know who has used BA as a DSL provider considers their attitude towards their DSL customers as a kind of uninformed apathy.
If nothing else they're doing a really terrific job of illustrating the very worst aspects of a monopoly.
jim
jim frost
Maybe you should converse with some of the people doing this work. There aren't too many "uneducated ... hackers" working on it so far as I can tell. Virtually every Linux hacker I know has formal education and works with Linux as a hobby. (And since it's a hobby they're more inclined to do it right than to ship something by June 1st, if you get my drift.)
In the end, though, who cares who wrote it so long as it gets the job done? I mean, you're assuming that the guys who put NT together were well educated, that any education they had was actually useful, and also that even if both are true they'll do a good job. That's a lot of very questionable assumptions, particularly about employees who are known to have built MFC and that brain-dead FIFO page replacement algorithm used in NT. (Ok, the latter made a certain amount of sense on the VAX. But couldn't they have done something better now that they have page reference bits?)
Now, bunches of NT people are sitting there thinking that I'm one of those Linux loonies, and there's certainly some bias on my part towards UNIX, but I was writing articles for UNIX people about the viability of NT back when NT was considered a joke by pretty much everyone in the business. NT is a damn fine workstation OS, particularly in a world where the majority of software is written for Windows, and I still use and recommend it in a lot of situations. Back in 1994 I figured that it'd decimate the UNIX workstation market -- and it did.
So NT isn't poison to me, but I have serious reservations about it as a server, most particularly because its stability isn't so hot, but also because I haven't been all that fond of how much I've had to spend to get extra software to do things that UNIX has done out of the box for years.
Yea, yea, I hear you saying "Win2K fixes the stability problem." So Microsoft claims, and maybe it's even true, but given what it's doing its hardware requirements are a little out of hand and the cost ... well, I think Microsoft is taking a lot of people for a ride. For a lot of server functions you can buy the OS and hardware for less than Win2K alone.
Then again, I'm educated enough to realize that I have a choice in the matter, and perhaps that's the real threat of Linux. It's not the 14 year olds, it's the guys who are smart enough to be comfortable with not using Windows. There are a lot of those guys, both old-timers and recent graduates, in IT shops and software development houses. Those are the guys who made Linux grow so much in the server space last year.
Maybe Linux really is made largely by 14 year olds and I've just not run into them. So what if it is? It's cheap, it's stable, and it has a hell of a lot of functionality. It's not always the best choice for the job, but you're stupid not to at least consider it.
Similarly, sometimes you have to bend over backwards to get Linux to do the job. This is particularly the case for a lot of specialized applications. So look around you and see what works best for what you have to do.
I suspect that for a lot of people that'll be a mix of OSs. It certainly is for me.
jim frost
I'd like to add one additional request in the name of bandwidth reduction. Display lists. One of the more interesting things about NeWS (Sun's post-SunView windowing system based on PostScript) was its ability to be programmed. You could create a routine that could render a whole button with bevels and shading and whatnot as a single function call whereas in X it will typically take six or more protocol requests to get the same job done.
While full programmability ala NeWS is rife with potential problems, given just the ability to record a sequence of parameterized protocol requests and request playback would drastically reduce the bandwidth requirements and boost the performance of a lot of applications while having very little impact on memory use. I believe display lists were considered and discarded originally due to the requirement that X11 run in a terminal with very little memory (too little to hold much in the way of display lists). I've never appreciated that since the client-side fallback technique was always a viable option, but given the opportunity to extend the protocol I think this would be something to consider.
jim frost
Sun marketing be damned, the main reason to use Java is not portability, it's development time. Versus C or C++ my productivity at least tripled and for certain kinds of applications (network servers in particular) it's way, way better than that. And that's been pretty much the same story for everyone else I know using it, too. Furthermore the bug counts are way, way down and the bugs tend to be irritating rather than debilitating (ie throw an exception and keep running rather than dumping core).
Swing and AWT could use a lot more maturity but if you don't need a GUI, or you only need a basic GUI, Java is awfully nice for building applications.
More speed I can definitely use but realistically speaking we're building some damn fast servers with Java, stuff that runs rings around the C++ competition, even using the Sun 1.1 JDKs. When you're building the code three times as fast you have a lot more time for optimization!
jim frost
They didn't help. Mostly, they were ignored.
If we're going to use computers pervasively in the classroom we need to use them as information resources -- modern textbooks -- not full-blown computational systems. It's just not cost-effective to give general-purpose systems to everyone.
When I was in high school in the early 80s the history books were printed during the Korean War. I didn't learn about Watergate until I was a senior (1984) and it's not like I was taking the basketweaving classes. Twenty year old textbooks were the norm! They couldn't afford to replace the textbooks more often than that.
Given that this is the case, any electronics in the classroom have to have economic benefit or it will simply not happen on a broad scale.
It's not real hard for me to see some possible economic incentive for giving everyone a $50 electronic book and having them download their textbooks. With publisher subsidies (and there are some good economic reasons why they might be willing to subsidize) these devices could break the $20 mark at the consumer level within five years. There are already workable readers selling for under $100 today so I think this is very achievable.
Since most of the cost of a textbook is in printing and shipping the end cost of outfitting a student with all the necessary texts would drop by an order of magnitude, enough that they could have up-to-date textbooks and still save the school money. And the publishers would have better margins, too.
Now, I can believe that there would be some value to the flexibility of a laptop or other general-purpose unit, but let's be realistic here. Public schools are driven primarily by cost and secondarily by benefit. If information technologies can reduce costs as well as provide benefit they'll go with them; if not, they'll keep using ancient texts.
jim frost
Given the broadly available connection technology you are correct, DVDs are not really instantly distributable. But he has a point: it won't stay that way forever. It probably won't stay that way for more than ten years.
What they're doing is trying to forestall the problem in the future. This is forward-thinking and exactly what they need to do to protect their businesses. It's really hard to have a beef with that.
Where they're missing the boat is that they can't stop the technology from being published no matter what they do. There are just too many jurisdictions and everyone with a connection is a publisher. It's a battle they lost the day DeCSS was published. Maybe if it had happened earlier in the life of DVD players they could do something about it, but now they have the choice of wimpering about it or making everyone buy new DVD players. The latter won't happen now that DVDs have exceeded critical mass.
Rather, they're going to need to come up with approaches to track down the offenders, just like they do for existing analog piracy. Nothing new here.
Now, he's right that digital distribution is going to make it a lot easier for the thieves to distribute their stuff. But -- and this is really important -- it's going to make it easier for the content producers to distribute it too. The cost of distribution is falling through the floor, and economies of scale will still work in their favor just as they did with videotape.
They're still going to make money in a world where instantaneous distribution is possible. They just have to find ways to make it worth the consumer's while to buy it rather than steal it, and by and large that means they have to keep it affordable.
jim frost
I don't agree with that approach either, but there's nothing stopping you from buying a cluster of smaller Suns (we recommend this approach) and since the base availability is much better (as is the manageability) it's a lot easier to keep running and you can use a much smaller cluster.
Having said that I prefer lots of smaller machines myself, but not running NT. You have to do too many things at the NT console, and reinstalling NT is really painful, and they just don't stay up as long as they should. A cluster of BSD machines is a rather cost-effective approach to this problem (isn't that what Yahoo and HotMail do?). At least you know that when one goes down it'll be because of a hardware failure, and you know you can reload the thing from scratch in about a half hour.
jim frost
In my experience (and what my company does is high-volume web applications) NT servers are really unreliable relative to Sun. It's not even a contest -- it's an order of magnitude difference. Furthermore, there are real limits to the size and power of a single NT machine relative to Solaris.
The article talks about serving 8.9M pages in a day as peak load. They don't tell you how many of these machines they're running, but we do better than that on one 4-CPU Sun 5000.
I find their harping on eBay pretty funny. Yea, eBay has had some really bad outages on Sun hardware, but that's not so much the fault of Sun as it's the fault of a really bad application architecture on the part of eBay. You're stupid if you think that you can depend on a single piece of hardware or software (no matter how good it is) never failing. It's something of a credit to Sun that eBay is as reliable as it is, and a credit to Sun that they were able to scale their architecture to the degree they have.
Our stuff runs dandy on both NT and Solaris, but customer feedback indicates that Solaris has vastly better reliability.
jim frost
Certainly, especially on the first iteration. (After all, you already have that monster box.) But it isn't always a Sun, and that means eroding market share. And even when it is a cluster of Suns you have to buy quite a lot of them before Sun makes as much money as when you buy a single Starfire.
Smaller boxes are bad for Sun. They're bad because you can buy them from some other vendor, and they're bad because they have smaller profit margins. And, like it or not, smaller and cheaper is the natural progression in this business.
jim frost