Weekly Edition
Return to the Press page
| |||||||||||||
|
| |||||||||||||
Three things the Linux desktops needs to do to beat Windows (ComputerWorld)
Steven J. Vaughan-Nichols covers
two LinuxWorld panels. "While at LinuxWorld at the Moscone Center in
San Francisco, I chaired the panel on what the OEMs (original equipment
manufacturers) that are pre-installing Linux on their PCs are up to and I
attended another panel on what the Linux desktop architects have
planned. One theme that showed up at both functions is: "What does Linux
need to do to compete more successfully on the desktop?" We came up with
several pain points, but some of them are clearly hurting Linux more than
the others. "
(Log in to post comments)
Better support for Windows applications and drivers - ugh Posted Aug 15, 2008 20:05 UTC (Fri) by khim (guest, #9252) [Link] OEMs are the ones who can help situation with drivers (I applaud Dell for their efforts) but "better WINE" is just stupid. WINE is acceptable if it's used by application MAKERS (like Picasa for Linux or some games) - they have access to both WINE sources and their application sources and can tweak both to find perfect match (and then bundle appropriate version of WINE, of course), but when application makers are actively trying to hurt WINE (like Microsoft)... forget about these applications - or else go back for Windows. Sorry. No way to fix that.
A niche for WINE Posted Aug 15, 2008 23:26 UTC (Fri) by dmarti (subscriber, #11625) [Link] There's a lot of old, infrequently-released, but hard to replace Windows software out there. Surprisingly at a lot of companies the deal-breaker for desktop Linux isn't something big like Adobe Photoshop, but some old application that they originally installed on Windows NT 4.0 or something and have kept going ever since.Add a solid WINE to the migrator's toolbox and you might succeed at some of those companies.
A niche for WINE Posted Aug 16, 2008 0:19 UTC (Sat) by sbergman27 (guest, #10767) [Link] Agreed. Unfortunately, its usually some old one-off VB app with a jet backend, and that is where Wine seems to have the most trouble. I always get a sort of "Abandon all hope, ye who enter here" feeling when I run into those situations.
Wine won't work for... Posted Aug 16, 2008 1:10 UTC (Sat) by pr1268 (subscriber, #24648) [Link] There's a lot of old, infrequently-released, but hard to replace Windows software out there. Surprisingly at a lot of companies the deal-breaker for desktop Linux isn't something big like Adobe Photoshop, but some old application that they originally installed on Windows NT 4.0 or something and have kept going ever since. A sysadmin friend recently lamented that his company can't wean itself off of SCO servers for part of their business because of a single 3rd-party proprietary application that they use. He told me that when he asked the vendor if/when they would release their product for Linux (or even Solaris), he was rebuffed with a reply similar to, "SCO is the only Unix there is." I've seen a lot of ISVs make related silly claims about Microsoft Windows, but I was certainly surprised to hear that at least one vendor feels the same way about SCO! Strange...
Wine won't work for... Posted Aug 16, 2008 11:49 UTC (Sat) by tialaramex (subscriber, #21167) [Link] This sort of small specialist ISV tends to see porting to Linux as an expense they'd rather avoid rather than as an opportunity to make their software more attractive. And they may even be right. If you sell the only software in the world for managing an inventory of Boeing commercial jet parts then logically every single company that could have a use for your software (that is, companies which operate and maintain Boeing passenger jets or the freight variants) already has it, and they buy whatever hardware and software you tell them is necessary. So porting to Linux doesn't mean more customers, it only means some of your existing customers are happier. If they won't quantify that additional happiness in terms of hard currency, why should you care ? But as an end user, at least in 2008 you don't need to manage and maintain hardware for all these vertical applications, because most of them can be virtualised inside managed RHEL machines. So there's no risk of embarrassing situations where you have to explain to the board that the fan broke on a ten year old PC under someone's desk running even more ancient DOS software that turns out to be essential to the minute-by-minute operation of your airline, costing $3 million in compensation to passengers.
Three things the Linux desktops needs to do to beat Windows (ComputerWorld) Posted Aug 16, 2008 1:29 UTC (Sat) by N0NB (guest, #3407) [Link] Desktop Linux already beats Windows hands down and has for some time. The proof is that now only three areas now seemed worthy of mention. Is that a simplistic assessment? Perhaps, but for the mass majority of desktops, a Linux based desktop will extend the life of the hardware and allow users to be comfortable in their computing environment for a long time. The real challenge is overcoming "one size fits all" IT department mandates that don't take into account the needs of individual customers. A smart and agile IT department will find plenty of places where a Linux based desktop can fill a role.
Three things the Linux desktops needs to do to beat Windows (ComputerWorld) Posted Aug 16, 2008 1:51 UTC (Sat) by vonbrand (subscriber, #4458) [Link] Having been a bit involved in "IT Department" style settings, I do fully understand the "one size fits all" philosophy: It is a lot easier/cheaper to support one standard configuration than having to find out anew each time what exactly the local hacks are and if/how they contribute to the present weird sympthoms. Also, having similar setups for all users allows them to help each other out, lessening the help desk burden. Don't get me wrong, I do love the flexibility open source gives, but I also know first hand how wannabe "experts in configuring systems" screw things up (and then vanish into the woodwork).
Three things the Linux desktops needs to do to beat Windows (ComputerWorld) Posted Aug 16, 2008 6:24 UTC (Sat) by dkite (guest, #4577) [Link] Yes, only three. Just minor stuff like full featured software. And working hardware. What has happened to make the Linux desktop a possibility is the dumbing down of the desktop. Much of what used to be provided by rich applications is now web based. Derek
Three things the Linux desktops needs to do to beat Windows (ComputerWorld) Posted Aug 16, 2008 15:41 UTC (Sat) by sylware (guest, #35259) [Link] All that is BS. My Dad and Auntie thanked me to have freed them from the evil OS, 2 years ago. About the desktop, they say it's quite easier to use than the evil OS once you are used to it, and I talked to other people who say that it's even easier than the fruit OS. That's the feedback I get. Now, the open source desktop has reached the maturity of the evil closed and non free one, it will start to really crush it since... it won't stop maturing and improving (like the Linux kernel did and is doing). The only members of the family still using the evil OS are my bros: they want to play the latest games, and most of them do not run on GNU/Linux. Nethertheless, most of the majors FPS has a GNU/Linux native version. We have wine but its use is way not easy enought for gamers, and many games have issues running. The UT3 client is still on hold because the evil company maneuvered to try to buy activision and managed to delay/cancel the GNU/Linux 32 bits and 64 bits clients. Rumors say that the WOW client is used internally and was never released... The *real* issue is most of the personnal computers around this planet are pre-installed with the evil one. The *only* way is to have the choice of the OS on 100% of the personnal computers sold around the globe. That is the *only* way out of the awfull state the computer industry is into. BTW, since the evil one is now copying Linux, we should do a GPL check on their code.
Three things the Linux desktops needs to do to beat Windows (ComputerWorld) Posted Aug 17, 2008 12:13 UTC (Sun) by BackSeat (subscriber, #1886) [Link] My Dad and Auntie thanked me to have freed them from the evil OSThis kind of attitude does Linux no favours whatsoever. Referring to an operating system as "the evil OS" is little short of stupid: for a start, it's unlikely that the OS itself is "evil", although you may have that opinion of the company that sells it. Even if you do, however, you'll convert few to your point of view with such language. People will either already agree with you (no conversion), or they will dismiss your comments as unblanced, biased, prejudiced and no doubt a handful of other adjectives. Allow me a small assumption. The company is Microsoft: it isn't "M$" or "microshaft" or "the evil one". For all the faults perceived in Microsoft and their products, they don't refer to Linux as "the evil OS" (on a public forum at least) - and, conceivably, they may have good reason to think of Linux in those terms. Show some respect for the other side. Rise above, don't sink below. (And no, I'm not a Microsoft fan by any means). BS
Three things the Linux desktops needs to do to beat Windows (ComputerWorld) Posted Aug 17, 2008 13:05 UTC (Sun) by zotz (guest, #26117) [Link] "For all the faults perceived in Microsoft and their products, they don't refer to Linux as "the evil OS" (on a public forum at least)" OK... What was the term I heard attributed to them? Cancer? drew
Three things the Linux desktops needs to do to beat Windows (ComputerWorld) Posted Aug 18, 2008 7:56 UTC (Mon) by Seegras (subscriber, #20463) [Link] Of course there is no "evil OS". An operating system can't actually be evil since it has no will of its own. It's the "bad OS".
an OS isn't evil (and don't label it as such) Posted Aug 18, 2008 21:31 UTC (Mon) by pr1268 (subscriber, #24648) [Link] This kind of attitude does Linux no favours whatsoever. Referring to an operating system as "the evil OS" is little short of stupid: for a start, it's unlikely that the OS itself is "evil", although you may have that opinion of the company that sells it. Well said--I agree with your advice. However, I must confess to having been amongst the name-callers you mention. But lately, I've been on a kick to hold back expressing my disdain for the proprietary OSes and put matters in perspective when comparing the merits and faults of various competing computer OSes. Doing so, I've gained an understanding (if not respect) of Microsoft's and Apple's way of doing things. Linux will succeed based on technical superiority and a rapid, continuous evolutionary development model. Also, I'd like to add that Vaughan-Nichols's three things could be reduced to two when considering desktops because (IMO) power management is less an issue with desktop workstations and servers.
power management Posted Aug 19, 2008 1:06 UTC (Tue) by andrel (subscriber, #5166) [Link] Power management is a huge issue with servers. Here's a back of the envelope:
That's just one server! If you're in 120V-land you can use a Killawatt to determine if 400 watts is a good estimate for how much your machines pull. (It is for ours.)
power management Posted Aug 19, 2008 1:26 UTC (Tue) by pr1268 (subscriber, #24648) [Link] I was referring mainly to suspend-to-disk and hibernation--two items I don't ordinarily associate with desktop and server computers. But, if Vaughan-Nichols's and your discussions are referring also to processor states and disk spin-down, etc., well then maybe I'm just a little uninformed or ignorant of how power management issues affect Linux. My intuitive thinking tells me that a lot of power management can be managed outside of the OS in the architecture and hardware engineering. Yes, I am aware that power consumption in the server room is a critical factor--most data centers charge monthly fees by the kilowatt-hour--but I just don't see how Linux is deficient in this area compared to other proprietary OSes.
power management Posted Aug 21, 2008 17:39 UTC (Thu) by mjg59 (subscriber, #23239) [Link] The majority of power management functionality is under OS control, not firmware. In any case, plenty of people in the desktop and server markets are interested in supporting suspend/resume.
power management Posted Aug 19, 2008 8:08 UTC (Tue) by ekj (subscriber, #1524) [Link] $526/year in power. For a machine that likely cost $10.000 to buy, has a useful service-life of no more than 5 years, and cost $10.000/year to administer and run. It's an issue, sure. If you could halve the power-budget WITHOUT compromising significantly on other things, it'd be worthwhile. But over 5 years the machine will cost a total of ~$60K, and it'll consume power for $2500. It's not as if it's the main cost. Things change if you have large numbers of identical servers so maintenance-cost pro server falls. But even if maintenance-costs where zero, power would still be only a fifth of the operating-expenses. (and we all know maintenance is never zero)
power management Posted Aug 21, 2008 14:29 UTC (Thu) by zlynx (subscriber, #2285) [Link] Yeah. And don't forget that for every watt used in a computer room, more watts have to be used to move that heat outside the building.
|
|
||||||||||||