Okay, I'm not being fair -- WAS: Oh joy, ZD ...
on
CNET Buys Ziff-Davis
·
· Score: 2
I've only read two articles since SmartReseller became SmartPartner and by the simpliest laws of statistics that's not being fair. My appologies as I do not read much ZD anymore.
But when I did, SmartReseller (now SmartPartner) was the only ZD publication I could stand. Unlike other ZD publications who seemingly play politics, SmartReseller catered to the OEM, integrator or solution provider who needed the job done while keeping his margins in the profitable range.
In such cases, SR did an excellent job (in most cases at least) of giving Linux a fair review.
-- Bryan "TheBS" Smith
Oh joy, ZD goes from pitfully ignorant to ignorant
on
CNET Buys Ziff-Davis
·
· Score: 3
Seriously now, at least one good thing comes out of this if ZD adopts c|Net's attitude, the freak'n article titles aren't so "sensational". I am personally sick of seeing an article's title only to read it and come away feeling the exact opposite.
From articles on Windows 2000 to Linux, it seems that the titles are for CIOs who don't bother to read the article. As such, they continue to use Microsoft and Microsoft partners' products.
Sm@rtReseller used to be a tolerable ZD publication, but ever since they switched their name to Sm@rtPartner, they might as well have switched their name to MicrosoftPartner!
Yes! That is what the UDF filesystem is all about. Note when I said above about just putting a line in the/etc/fstab file. It's just like another mounted partition.
I've been people talking about DVD-RAM in this thread who've never seen a drive nor the media (so I explain it below). Also, there is this "mis-nomer" that Windows actually has built-in writable UDF support -- NOT!!!
The DVD-RAM drive loading mechanism is quite ingenous as it supports both cartridge and plain disc! It has a spring-loaded tabs on the ends with slits so plain disc media (like single-sided DVD-RAM disks) slide in much like a slot CD/DVD drive (although much more nicely), and the tabs move back when the cartridges are used. BTW, the cartridge format is mechanically exact to the old CD caddy form-factor, but it is keyed so you cannot use a caddy (only half of my old CD caddy from my Plextor 6Plex will insert before the key hits).
This "unified media" tray only comes out about 1.5" (4cm) so it won't break off either (at least not easily), and it does an excellent job of autoclosing when you've "push the disc/cartridge in enough" (a very nice touch that still works after 18 months of hard use as my main CD drive as well). Personally, I'd like to see this type of mechanism adopted in all CD/DVD drives (at least in the standard 1U/half-heigh drives on desktops). You have to see this in action to appreciate the elegant design.
Not even Windows 98 SE has writable/re-writable UDF support, only read-only UDF support (and apparently limited at that?). You must use additional software from Software Architects to get UDF support in Windows 95/98, or use another program like Adaptec DirectCD to write directly to CD-RW drives (not sure if Adaptec supports DVD-RAM yet). This is an important consideration if you want to buy a DVD-RAM drive for use with Windows 9x, because SAI's UDF will cost you $79 if you buy the bare drive without it (whereas some kits like Creative's bundles SAI's UDF for Windows 9x). Also understand that SAI sells Windows 9x drivers separate from the Windows NT/2000, so it'll cost you double for both Windows 9x and NT/2000 support ($79/each). Makes you appreciate Ben's open source UDF driver for Linux even more!;->
BTW, I have heard that Windows 2000 does not even have read-only UDF support so it will NOT read regular DVD-ROM data disks? (someone please confirm?)
[ Please moderate this up because I am using it right now on Linux, and have been for almost 6 months! ]
I have been using my Panasonic/Creative DVD-RAM drive for almost 6 months now under Linux (and have had the drive for ~18 months). RedHat kernels since 2.2.12 have detected it and installed a SCSI generic disk driver (as/dev/sda since I have IDE drives). The reason why DVD-RAM was supported so quickly in Linux is because much of the firmware is similar to the old Panasonic PD drives (remember, rewritable CD before CD-RW?;-).
Ext2 works fine on it if you decide to format it. Otherwise, a simple download and compile of Ben Fennema's UDF driver (no complicated kernel patch necessary, just./config, make, make install installs the VFS module necessary) and you're cooking with an OS independent filesystem on media that lasts 30+ years! 2.6GB per side (with newer 4.7GB drives/media as the above pointed out). Again, it's simple. Just put the "/dev/sda" line in your/etc/fstab as normal with "udf" as the filesystem (assuming you've done the above). I assume you can do similar with CD-RW drives and the UDF driver as well (SCSI CD-RW drives at least).
For those of you not familiar with rewritable DVD, there are various formats. DVD-RAM was supposed to be the "standard." Of course that didn't stop Sony, Philips and others from breaking away from Panasonic, Creative, Matsushita, Pioneer and others to create their own, proprietary standards. The reason I choose DVD-RAM is because unlike most other DVD drives (most of the DVD-ROM drives of the time, fall 1998) is because they had trouble reading CD-RW media, and even some 2nd generation drives had trouble reading CD-R media (DVD-RAM reads all media: CD-ROM, CD-R, CD-RW and CD-various formats). But understand that no rewritable DVD-RAM I know of allows you to burn DVD-R, nor even CD-R/RW (although it was rumored that Philips had a proprietary 3/6GB DVD-rewritable that could also CD burn/re-write as well? But I never saw it myself). Another reason why I went with DVD-RAM is because some 2nd and most 3rd generation DVD-ROM drives could read it physically (at least the non-cartridge, single-sided version) and non-Panasonic/Creative/Matsushuita/Pioneer drives only needed a firmware upgrade to do so. And DVD-RAM is rewritable at 1,350KBps (1x DVD, 9x CD) whereas many CD-RW (and even some other DVD-rewritable formats) are a measly 300-600KBps (2/4x CD).
Anyhoo, while other vendors talk about rewritable DVD sizes and capabilities, Panasonic delivered a long time ago. And now they are boosting the size to 4.7GB/side with the possibility of CD-RW compatibilty. You can get Panasonic 5.2GB DVD-RAM drives for $200-250 nowdays (and I only paid $500 for mine in fall of 1998), with the 2.6/5.2 single/double-sided media for $20/30, respectively. It's not hard disk speeds, but it is massive storage at cheap prices. With Pioneer and others finally giving Panasonic/DVD-RAM a boost in portable video equipment within the last 12 months, I'd say DVD-RAM will become the standard that it was originally spec'ed to be. With a 30+ year shelf life, it's a great archiving format for 10+ years where magnetic tape is not. And unlike other optical formats, DVD-RAM is an open standard which means that future drives should be able to read it -- a very important factor when considering long-term archiving because who cares if it lasts if you won't have a drive that can read it!
DVD-RAM is great for video editing systems, for which, I bought my DVD-RAM drive to complement my brand new Matrox Marvel G200-TV at the time. Again, much, much cheaper than magnetic disks per MB/GB.
AT&T's Virtual Network Computing (VNC) is an excellent GPL licesned product to look at. When the server is run on UNIX systems, it allows you to run upto 99 independent sessions (more if you reconfigure/recompile it) on a single UNIX system. These sessions have their own window manager, programs, etc... tailored for the user running it. Essentially, you have a full X-Session running just like if the user was directly in front of the GUI console of the workstation. The user can then connect to the session from any VNC client, whether that client is MacOS, Windows, UNIX, DOS, Palm, etc...
Using VNC instead of a traditional X-Server on the PC side has several advantages. One is the fact that a user can disconnect and reconnect their viewer on another system and not have to logout. This also means that if the VNC viewer system (e.g., Windows) crashes, their entire X-Session is still running (and can be reconnected to). VNC also works fairly well over low-speed connections (as good as X11R6.3 extensions for low-speed connections), provided you minimize background images. An additional advantage is the fact that it runs on a single port (5900 + session #) which makes SSH tunneling extremely simple (side benefit: "low-cost, more secure" remote access than most "all ports open" commercial VPN software).
I work for a company whose applications are 90% UNIX based (Theseus Logic), and that's not likely going to change soon (as EDA tool vendors are choosing Linux over NT because of the true multiuser capabilities). We use Linux and Solaris sytems to run these applications. Although we are starting to dual-boot some of our NT Workstations on our desks with Linux (although my personal workstation and all our servers are 100% Linux;-), most of our work is done via VNC over to our headless Linux and Solaris systems. With 512MB to 1GB of RAM, we can easily accomodate 10 engineers on each system with fairly intensive engineering applications running. This has additional benefits like accomodating node-locked licenses that normally won't remote display (to another system), but will work in a VNC session (because the VNC session appears simply as display "localhost:vncsession#.0").
VNC is also a great way to slowly move to Linux. Users can spend 6 months becoming familiar with Linux use via a remote VNC session, while still running Windows on their desktop. You can also use VNC to UNIX systems so any user can run those few (or many, like us;-) UNIX applications when needed. Again, VNC is so simplistically powerful (especially on UNIX systems), you'll never run out of uses for it.
Again, instead of trying to deal with finding a costly terminal server program, or a PC-X-Server, evalute if you really need one. If your UNIX platform supports VNC (and I seriously doubt there is a major or even minor UNIX platform where it has not been ported to), just give your users a full UNIX session with all it's goodies. Best of all, with the VNC server and all its windows/apps running on the UNIX platform itself, you have 100% native execution.
[ Side Note: Unfortunately, since Windows is not a true multiuser system (and only a bastardized one with products like Terminal Server), you cannot use the VNC server in a reversed role (multiple users on a Windows NT system). Although you can use VNC like pcAnywhere (unified single session, remote and local user both see the same and control the keyboard/mouse simultaneously). And like pcAnywhere, VNC sessions can be shared, allowing remote training, etc... (even on UNIX, with multiple sessions running) ]
I think people realize that Linux is both a consumer platform and a business/server platform. The Windows 98 / Windows NT-2000 platforms are distinctly separate platforms, with each only filling in one of those. Comparing Linux to Windows NT-2000 changes the game radically.
And don't count out forthcoming distros with kernel 2.4.x and XFree 4.x. I think it will level the playing field.
The argument is clear-cut and to the point. The web was designed to be a platform-independent repository of information and services. The W3C has done an outstanding job to give commercial vendors the tools and standards to even implement extensions to them for vendor-specific documents and capabilities through technologies like XML and DOM. These technologies were designed specifically so vendors could release new technologies and not have to wait on the "slow" standardization process to catch up. For Microsoft not to use these well-defined and documented technologies to implement almost any equivalent Win32 capability is to say they do not believe in a vendor-neutral standardization process, but believe in a Microsoft-generated standardization process.
And that process is the reason why no API lasts more than 2 years in the Windows world, no program is stable for more than 2 days in the Windows world and no independent thought exists in the Windows world more than 2 seconds. And this is coming from a former Win32 developer and enthusiast.
I trust Borland. They basically pioneered high quality development tools for the PC, while also pioneering the like a book mentality to software licensing.
Understand this isn't Microsoft we're dealing with, but Borland. I trust them to license commercial software and they make high quality software at that. With their powerful Object Pascal language and Delphi system behind a number of quality Windows applications (from largely non-Microsoft partners), it will be nice to get those applications ported to Windows.
And this is coming from a guy who prefers GPL, GNU/FSF, GCC, GTK+ and Gnome itself!;->>>
I'm a longtime OS/2 fan and love the Workplace Shell. It's a full-up, truely object-oriented desktop (unlike Microsoft's Active Desktop that only has some O-O-like attributes). I've seen some OS/2 WS-like programs for X, but I'd like to see a full window manager like it. IBM, where's the source? [ wishing too much I guess;->>> ]
Even without a OSS WS-like environment, can we at least get a file-oriented window manager? Personally, I'd love this. I hate just about everything about Windows with exception that the window manager is file-oriented. I guess that is what the Gnome 2.0 focus is on, but I'd really just like a light-weight window manager that was file manager-based -- even if you had to make some footprint decisions to leave out support for some graphics types, etc... I personally think it's overdue for X.
Anyone agree with me here? I like the CLI (which is *NOT* really a bad thing in UNIX because the shell capabilies are 100x better than Windows, hence why I always install Cygwin/BASH on NT;-), but that 2% of the time I wish I had a file-oriented window manager.
I've had issues with older PCI sound cards on newer PCI 2.2-spec mainboards. Nothing new, nothing limited to AMD either.
As far as the FPU bug, it goes in the same basket as the 100+ other errata on any Intel chip. Every mainstream microprocessor has bugs, dozens of them. Heck, my company just released a 8-bit microcontroller and, not even back from the Tawainese fab yet, we've already discovered our first errata (a small one that is easily corrected in firmware/software)! But Intel seems to take the cake and it's not just because they're #1 either.
On with the FPU thing, I've seen some IGNORANT people get on the Quake benchmark BS. I'm sorry, but idSoftware's Quake engine uses integers for a lot of things -- but because the Intel ALU is so freak'n slow (1/3rd the speed of any K6 chip), idSoftware just found that loading integers via the FPU load instruction on the Pentium ended up being faster (even though it was some 4 instructions to 1 do to the same thing -- talk about a design flaw!). As such, even though the K6 would have actually been faster than with the original code, the "Pentium optimized" version results in the crippling the advantages of the K6 in order to better optimize an otherwise poorly designed Pentium.
Remember, "Pentium optimized" means that this is how you must write your software to fix Intel design flaws in the Pentium (at least 25-30% of the time). Intel builds full-custom ICs which means you get the human error factor (although it does have some advantages), compared to many other firms where a lot of design is done by optimizers and other EDA tools.
This is just like the whole problem Gateway 2000 had with the AMD K6-2 systems -- it is NOT the chip -- but the selection of components around it. With the AMD K6-2 systems, Gateway 2000 choose a sub-standard mainboard (cannot remember the vendor), largely based on cost in volume (and possibly guaranteed volume as well). Now they have what looks like to be a current draw issue that is due to the mainboard or power supply. I mean, did c|Net even bother to check if AMD is using an Athlon-certified power supply???
I personally think AMD's track record as of late is better than Intels! Lest we forget Intel had a mainboard recall on the original SE440BX (reference i440BX chipset) mainboard due to power supplies frying the mainboard and its components. And I'm not going to go into the whole MTH (memory translator hub) fiasco and the RDRAM RIMM signal integrity that preceded it. I really would like to know what mainboard Gateway 2000 is using for this system. If its past K6-2 problem is an indicator, I'd say it's Gateway 2000 going for the lowest bidder.
Which makes me wonder why they pay a premium for their PIII mainboards instead of using someone cheaper than Intel??? Let's see here, for Intel, use costly Intel mainboards instead of cheaper alternatives (unlike others like Micron who uses the more flexible, VIA-powered Tyan S1854 Trinity 400)... but for AMD, use lowest bidder instead of paying extra for AMD stock... hmmm... First the K6-2, now the Thunderbird, that's STRIKE TWO GATEWAY 2000! [ Or was this intentional??? ]
I think people fail to realize that NASA has been fairly responsible in their space endeavors. For example, NASA only uses decaying radioactive materials in spacecraft that do not orbit Earth, and only when the power requirements of the craft deem it absolutely necessary (i.e. it is impossible to power it with solar panels due to solar distance and weight). NASA even went as so far as to worry about the tiny ammount of radioactives on the Lunar Module of Apollo 13, as to plot a much more risky (to the crew) return trajectory (with additional corrections) so as to not have the LIM crash into anything but a remote and deep watery grave (see "From the Earth to the Moon: Episode 8" to hear more on "the Nuke Problem" as the media called it).
This is in stark contrast to numerous Russian Earth orbitals, including even purely scientific ones! Take note of a Russian scientific satellite that contaminated hundreds of square miles (600 mi^2 I believe) of Canadian national forest when its return trajectory was not accomodated correctly. And I need not mention that it is still up in the air whether or not the Russian military, let alone our own, uses radioactives in Earth orbiting platforms.
I work at a chip design firm (Theseus Logic) and 90% of our applications run on Linux or Solaris all from a single Linux server (although we looking to add a 2nd server or a NetApp box). Outside of those programs used by 30-35 engineers, I spend 75% (or more) of my time messing with stupid Windows applications for the admin staff, a measly 7 people (even though the Engineers are 90% of the traffic and data).
From a $30K accounting package (Deltek Advantage) with its own NT server that cost more than our Linux box (and may require a 2nd one soon for stupid Citrix Winframe), to stupid little $2-5K/each Windows software packages here and their for inventory, stock options, etc..., I'm going up the wall. Especially when updating software (never goes right, unlike our UNIX EDA and other tools) and I pull my hair out. Everytime I bring up ERP I get told that since we've already spent >$50K plus another $50K on consultants, so we're not changing. Of course I brought up the point before we spent this money so I get the underlying "crying over spilt milk" or "quit rubbing it in my face" attitudes nowdays.
I can argue TOC with ERP, but for companies like mine that have already spent >$100K on disseparate Windows packages and don't want to pay anymore, a free/OSS package is the only way to get it in house. I sure wish companies would realize that maintaining disseperate little (and even big) Windows programs are just a pain in the @$$. I'm sorry but all it takes is 6 months of UNIX administration and sysadmins realize that UNIX maintainance is just 10x easier (thank God 90% of our engineering apps run on UNIX).
Thank God projects like GNU Enterprise and the Java-based Kontor Project have sprung up. I'd say if you want to help Linux get inside corporate America, look to donating your time on these projects. And you don't have to even be a developer to do so, I'm sure both projects are looking for a lot of bookeepers and accountants for most of the design.
You no longer have to buy a proprietary x86 OEM solution to get massive, multiple memory and PCI busses for high I/O throughput. The new ServerWorks (formerly Reliant Computer Corporation) ServerSet III-series chipsets are making their way onto retail mainboards from major vendors. They put the disk controller on a different PCI bus than the Ethernet controller.
For those interested, here is a comparison of various chipsets and their aggregate memory and PCI throughput in MBps, respectively (sorry about the format but./ doesn't seem to allow tables or the pre/code tag):
Again, this is just aggregate throughputs. RDRAM is faster at bursts, but SDRAM has lower latency (i.e. better), and the old EDO of the i450NX chipset doesn't even get close to what SDRAM and RDRAM can do.
But note the massive PCI throughput of the ServerSet chipsets from ServerWorks -- due to their 3 independent PCI busses, of which, 2 are 64-bit (both are 66MHz capable in the HE/WS). They're so good, that even Intel is adopting the 4-way IIIHE for a forthcoming SDRAM server mainboard instead of RDRAM. This is largely because the 4 RIMM slot i840 can only support upto 2GB of RDRAM whereas the 16 DIMM IIIHE can support 16GB. As of right now, all Intel can offer for servers is the 3-year old i450NX so most tier-1 vendors have opted to work with ServerWorks instead.
First off, it's really up to the programmers whether or not formalize their organization and follow the two and even third rigorously. I think he has made an analysis that all Open Source development is done in an unorganized, non-traditional software engineering environment and all commercial development is.
But let's say for a moment his assumption is correct. This brings me to the last two phases.
In stage three, commercial interests can drastically cut short the testing phase. The "rush-to-market" attitude prevails as software quality takes a back seat to profit margins -- the great conflict-of-interest that greatly reduces consumer benefit of the product. Vendor's release early way to often. A common misconception is the fact that while the "release" of an OSS product is less definable, because the source code has always been available, if software bugs are inherit to programs, why not always always make it available then? I think he fails to realize the testing _and_ recoding are inherit in use of an OSS product. And OSS projects still have to have the traditional "feature freeze" milestone to move even close to production quality. Unfortunately, commercial software cannot be recoded by consumers once released which makes it actually less tested than OSS -- again, especially if it falls victim to the "rush-to-market" mentality of most commercial software firms.
Then there's the final stage, which is 85% of the traditional software engineering model. Unfortunately, in today's shrink wrapped software world, it has been reduced to 5%. Why? Because there is basically no revenue stream in fixing commercial, shrink-wrapped software. As such, it does largely go unfixed, only being bothered with if a large volume customer complains enough and is paying several millions or tens of millions of dollars in support costs. OSS support never ceases because there is no set "release" of code, it's always been released. As such, OSS is in a continual state of support.
And that final stage is the biggest point. Again, the traditional software engineering model has always been about 85% post-release support. Only those design teams, whether they are OSS or commercial, who are driven by post-release support (either ethically or even financially) can accomodate this most important phase best. That's why when RedHat and other vendor's say "service-focused software firms" are the key to good software products, they are hitting the mark!
The days of rushing the 3rd phase and ignoring the 4th phase of the traditional software engineering model are over. OSS is here and it's our best chance at releasing quality software because it's inherit design methodology accomodates all phases of the software engineering model, even if it lacks an "official source code release" and the code can be chaotic at various points.
But even the concept and problem of the "choatic code state" is debatable. I mean, do you trust software just because it has been "officially released"? Or do you at least hear some peer-review before using it? I think the later applies to the adoption of all software more than the former.
Our Sunnyvale office tried to buy some Dell systems with Linux pre-installed. They kept getting the same answer over and over, Linux only supports 128MB of RAM (not the model, but Linux!). They could not even get the memory part number from the sales department so they could call the parts department to order more. It was ludicrious!
So our SV office made the mistake of ordering another model with Windows NT and ran into mega problems with everything from the video card to those sorry-@$$ Promise cards. Sure enough, someone finally contacted me (the order happened without my knowledge) and I tried to straighten out the whole fiasco (still working on it).
At our main office (where I work), we've dropped Dell completely because they have their head up their @$$ for the last 18 months -- seriously. I know they've had a lot of sales, which means they are overloaded and overworked in most departments, but Dell's recent expansion has left them in a cluster-f--- state, unlike the massive expansion of 1994 that Gateway 2000 underwent that was at least 3x as bad, but GW2K still kept good records and kept the order. After a year of Dell shipping systems to the wrong address (although billing to the right one, we couldn't even get it f---ing changed to be the same as the billing after 2 written letters!!!), shipping us replacements weeks late, and then shipping 2-3 of them (which only drives up their own costs!) and, lastly, because treated us like a f---ing 2-year old when our CEO's hard drive went belly up, with 3 different departments all pointing fingers at each other and no one to talk to, we dropped them. I mean, when you give my superiors some lame @$$ excuse time after time again, telling them flat out you don't know anything and can't do crap even though you purchase 6-figures of product a year, you drop a company like that. We gave Dell at least 2 dozen "chances" and they failed every single one... continuously.
I think everyone knows the moral of the story here: freak the direct PC OEMs. Stick with either established Linux vendors like VA or use some "real" enterprise PC OEMs like IBM, HP and the like. They don't pop hard drives in like bread in a toaster, they'll customize your solution.
For some reason, StarDivision (and now Sun) continue to assume that only one user will run StarOffice. While I have to put up with that kind of crap in the Windows world (e.g., most Windows software is ignorant of the multi-user issues of Windows NT/2000, including most of Microsoft's own software), you'd figure Sun would know better. As such, when I first installed StarOffice 4.0 a long time ago, I installed it as root and always had issues afterwards when running as other users (even when I opened up the permissions).
Of course, the multiuser installation is hidden with the command-line option "/net". Why this is either not the default or presented in the GUI as an option (at least they _do_ put it in the GUI readme now), I have no idea. When you install with the "/net" option, you install a shared copy that is just as good for multiple users on one system than multiple users over a network of systems. In fact, the "/net" label is really a misnomer, it should be "/multiuser" or something similiar.
The only other thing to remember is to install in another location than "/root/OfficeX" -- something that Sun should change when the "/net" option has been specified at the command line. While it is recommended that after this "multiuser install" each user run the./setup binary that is located in the shared StarOffice binary directory, it's completely unnecessary. Due to the power of UNIX's multiuser orientation, StarOffice (at least 5.x does) notices that you haven't done the ~2MB user-installation yet if you run the./soffice binary in the shared StarOffice directory. Simple no? [ Why not the default then?!?!?! ]
The last complaint is the continued political BS regarding the bundling of Java with StarOffice. If you have Windows, it is bundled. If you have Linux, it is not. [ Hmmm, have to check out Solaris and see if it is. ] That is just utter BS and seems to be a political issue with the Java license -- you can bundle it with a Windows binary product, but not a Linux binary product. Nuts!
Now on the bright side, one very cool thing about the new StarOffice 5.2 installer I just noticed is the lack of the requirement of the DISPLAY variable to be set correctly. E.g., I just su'ed in a terminal to root to install via the "/net" option and StarOffice figured out that it was running on the local system and set the DISPLAY variable correctly. Sure wish other vendor's installers would do the same. [ Now what happens if I try the "/net" install without being root? ]
Not only is there the standard waste in packaging, but what about the waste is shipping costs, store and shelf space before sale and the like? This level of unprecedented waste goes well beyond the "killing of trees".
You'd figure that resellers with a physical store front would have been complaining for years, let alone today's software warehouses. I mean, by reducing at least the 3rd dimension, depth, you can package 3-8x as many units while still preserving the marketing value of the 2-D frontal image! You don't have to be a rocket scientist to figure that one out! Dooh!
If you're like me, you restribute the CDs on a quality CD-R (e.g., I like Kodak replicator grade) with a nice label, ask $5 for it, and give out 2 freebies to new users for that 1 person that pays you. I only "ask" $5, but do not require it. As such, would that not be a "donation" and separate from the sale?
So I see it as this: I'm merely taking a donation and/or just trying to cover my costs (of which, I actually lose because at ~$1.66/each, the cost of the CD-R, label, color ink, printer, PC, etc... ends up being more). Again, I don't require people to pay for it either, but most people that are asking for their 2nd, 3rd+ CD feel they've already had their "free one" and spot my $5.
I also consider the distribution to be "local", kinda like if you were making copies for and to yourself. I just do it in my local LUG(s).
Lastly, I do download the source, just in case anyone wants it. It's available upon request. And maybe that is how you should approach it, "available upon request." Hmmm, need to read the GPL on that one.
While I'm not totally going to dismiss this guy's viewpoint, he should really try to explaining further some of his Linux "assumptions". They are "assumptions" because I they do little but prey upon the common myths about Linux. E.g., such as the fact that without a single GUI framework, no one likes it.
I think MacOS X will be a great addition to the OS landscape, especially if Apple keeps the codebase as close to FreeBSD as possible. But I don't see everyone going out and rushing to buy MacOS X unless it gathers some more and heavier hardware support on its own, single platform.
While the film had technical holes all over it, I found the underlying interactions between a conquering species and man to be interesting and consistent.
For example, the viewpoint of the aliens of human conversation, and vice-versa. The general distain and disgust, ignorance and indifference. Nothing breakingly original, but I noted on many occassions that the dialog and actions varied from the traditional interactions of two people (or species in this case) that could care less about each other. Again, I found it very consistent in many point, with the stupid technical issues interrupting it (to the point of my disgust!).
But what really drove me was all those technical impossibilities. Like the Harriers, bombs and computers that lasted 1000 years. The "learning machine" that thought the man to speak their alien lanaguage, but the aliens did bother to learn man's. Although that could be more of the interaction element I was talking about -- but still a stretch in application.
All-in-all, the movie would still have to be radically re-written to make it believable. Even if this meant that mankind "loses" -- it still would have been much, much better than with the general Hollywood ignorance and ending.
I concur with just about everyone here that the Linux Router Project (LRP) is a floppy solution that can run on even a lowly 386 CPU. You should be able to find such a system for $50, and not have to spend the $$$ you mentioned.
Otherwise, if you really don't want to use a PC, I'd grab something like the SonicWall SOHO/10 for around $400. As of last year, SonicWall's products were the only ICSA-certified firewalling products for under $4K. The SOHO/10 is a little 25MHz 68300-powered Coldfire running some RTOS (probably VxWorks). The SOHO/10 allows upto 10 nodes transparent access out, and even provides one-to-one NAT (private-to-public IP mapping) if you want to share out services, which you can filter, of course, by service.
Just FYI, their high-end product, the SonicWall PRO, is powered by a 233MHz SA 110 StrongArm chip and features a myrid of VPN and encryption options built-in, along with a DMZ port. It lists for $2995, not bad for its capabilities. But I figure you're not looking to spend THAT much.;->>>
Dude, I think most MacOS and Linux users and the applications for each systems are quite different.
To take your example, I would tell the home user to get a system with Corel Linux PRE-INSTALLED instead of trying to install it him or herself! And there are other issues (again, I didn't read the whole thing because it made me sick).
There was such a narrow focus (like the fact that you can get Corel pre-installed) and a lot was left out that it was just pitiful. Add to the fact that most home users would be better served by Windows instead just kills me.
Again, most MacOS and Linux users are quite different beasts, at least at this point. They should have never been compared as there is little overlap in their markets.
And I do not consider "anything but Windows" to be a good argument or market. Apparently this is what the media keeps harping on, and that makes Windows advocates hate Linux even more.
And yes, I use Linux everywhere, at home and work because I'm not even 25% as productive in Windows.
Guys, I'm sorry. I have never seen a more poorly written article on a comparison that should have never been made -- just exactly who was this written for?!?!?! I don't know what c|Net was thinking... maybe "Hey, let's compare the two non-Windows OSes that we think are easiest," but they even failed that.
I mean, the very first comparison, "installation," that was just rediculous!!! Could they at least find a PC vendor that pre-installs Corel? If not, then it is NOT a good comparison!!! Duh!
And who "wins" a category? Com'mon, you didn't even define how you choose. I mean, how did you weight hardware compatibility? If MacOS was written for Mac hardware (which means just about everything is supported), how did Corel win if it doesn't support all Intel hardware? What do you mean it's "less expensive"? Now you are jumping criteria! Please define them beforehand! So many laws of comparison were broken here.
And that's what this eventually boiled down to. Too many variables against extremely different platforms. Geez, just use the system that runs your apps.
I'm still trying to figure out who this article is written for?!?!?!
Slashdot Mirror
I've only read two articles since SmartReseller became SmartPartner and by the simpliest laws of statistics that's not being fair. My appologies as I do not read much ZD anymore.
But when I did, SmartReseller (now SmartPartner) was the only ZD publication I could stand. Unlike other ZD publications who seemingly play politics, SmartReseller catered to the OEM, integrator or solution provider who needed the job done while keeping his margins in the profitable range.
In such cases, SR did an excellent job (in most cases at least) of giving Linux a fair review.
-- Bryan "TheBS" Smith
Seriously now, at least one good thing comes out of this if ZD adopts c|Net's attitude, the freak'n article titles aren't so "sensational". I am personally sick of seeing an article's title only to read it and come away feeling the exact opposite.
From articles on Windows 2000 to Linux, it seems that the titles are for CIOs who don't bother to read the article. As such, they continue to use Microsoft and Microsoft partners' products.
Sm@rtReseller used to be a tolerable ZD publication, but ever since they switched their name to Sm@rtPartner, they might as well have switched their name to MicrosoftPartner!
-- Bryan "TheBS" Smith
Yes! That is what the UDF filesystem is all about. Note when I said above about just putting a line in the /etc/fstab file. It's just like another mounted partition.
-- Bryan "TheBS" Smith
I've been people talking about DVD-RAM in this thread who've never seen a drive nor the media (so I explain it below). Also, there is this "mis-nomer" that Windows actually has built-in writable UDF support -- NOT!!!
The DVD-RAM drive loading mechanism is quite ingenous as it supports both cartridge and plain disc! It has a spring-loaded tabs on the ends with slits so plain disc media (like single-sided DVD-RAM disks) slide in much like a slot CD/DVD drive (although much more nicely), and the tabs move back when the cartridges are used. BTW, the cartridge format is mechanically exact to the old CD caddy form-factor, but it is keyed so you cannot use a caddy (only half of my old CD caddy from my Plextor 6Plex will insert before the key hits).
This "unified media" tray only comes out about 1.5" (4cm) so it won't break off either (at least not easily), and it does an excellent job of autoclosing when you've "push the disc/cartridge in enough" (a very nice touch that still works after 18 months of hard use as my main CD drive as well). Personally, I'd like to see this type of mechanism adopted in all CD/DVD drives (at least in the standard 1U/half-heigh drives on desktops). You have to see this in action to appreciate the elegant design.
Not even Windows 98 SE has writable/re-writable UDF support, only read-only UDF support (and apparently limited at that?). You must use additional software from Software Architects to get UDF support in Windows 95/98, or use another program like Adaptec DirectCD to write directly to CD-RW drives (not sure if Adaptec supports DVD-RAM yet). This is an important consideration if you want to buy a DVD-RAM drive for use with Windows 9x, because SAI's UDF will cost you $79 if you buy the bare drive without it (whereas some kits like Creative's bundles SAI's UDF for Windows 9x). Also understand that SAI sells Windows 9x drivers separate from the Windows NT/2000, so it'll cost you double for both Windows 9x and NT/2000 support ($79/each). Makes you appreciate Ben's open source UDF driver for Linux even more! ;->
BTW, I have heard that Windows 2000 does not even have read-only UDF support so it will NOT read regular DVD-ROM data disks? (someone please confirm?)
-- Bryan "TheBS" Smith
[ Please moderate this up because I am using it right now on Linux, and have been for almost 6 months! ]
I have been using my Panasonic/Creative DVD-RAM drive for almost 6 months now under Linux (and have had the drive for ~18 months). RedHat kernels since 2.2.12 have detected it and installed a SCSI generic disk driver (as /dev/sda since I have IDE drives). The reason why DVD-RAM was supported so quickly in Linux is because much of the firmware is similar to the old Panasonic PD drives (remember, rewritable CD before CD-RW? ;-).
Ext2 works fine on it if you decide to format it. Otherwise, a simple download and compile of Ben Fennema's UDF driver (no complicated kernel patch necessary, just ./config, make, make install installs the VFS module necessary) and you're cooking with an OS independent filesystem on media that lasts 30+ years! 2.6GB per side (with newer 4.7GB drives/media as the above pointed out). Again, it's simple. Just put the "/dev/sda" line in your /etc/fstab as normal with "udf" as the filesystem (assuming you've done the above). I assume you can do similar with CD-RW drives and the UDF driver as well (SCSI CD-RW drives at least).
For those of you not familiar with rewritable DVD, there are various formats. DVD-RAM was supposed to be the "standard." Of course that didn't stop Sony, Philips and others from breaking away from Panasonic, Creative, Matsushita, Pioneer and others to create their own, proprietary standards. The reason I choose DVD-RAM is because unlike most other DVD drives (most of the DVD-ROM drives of the time, fall 1998) is because they had trouble reading CD-RW media, and even some 2nd generation drives had trouble reading CD-R media (DVD-RAM reads all media: CD-ROM, CD-R, CD-RW and CD-various formats). But understand that no rewritable DVD-RAM I know of allows you to burn DVD-R, nor even CD-R/RW (although it was rumored that Philips had a proprietary 3/6GB DVD-rewritable that could also CD burn/re-write as well? But I never saw it myself). Another reason why I went with DVD-RAM is because some 2nd and most 3rd generation DVD-ROM drives could read it physically (at least the non-cartridge, single-sided version) and non-Panasonic/Creative/Matsushuita/Pioneer drives only needed a firmware upgrade to do so. And DVD-RAM is rewritable at 1,350KBps (1x DVD, 9x CD) whereas many CD-RW (and even some other DVD-rewritable formats) are a measly 300-600KBps (2/4x CD).
Anyhoo, while other vendors talk about rewritable DVD sizes and capabilities, Panasonic delivered a long time ago. And now they are boosting the size to 4.7GB/side with the possibility of CD-RW compatibilty. You can get Panasonic 5.2GB DVD-RAM drives for $200-250 nowdays (and I only paid $500 for mine in fall of 1998), with the 2.6/5.2 single/double-sided media for $20/30, respectively. It's not hard disk speeds, but it is massive storage at cheap prices. With Pioneer and others finally giving Panasonic/DVD-RAM a boost in portable video equipment within the last 12 months, I'd say DVD-RAM will become the standard that it was originally spec'ed to be. With a 30+ year shelf life, it's a great archiving format for 10+ years where magnetic tape is not. And unlike other optical formats, DVD-RAM is an open standard which means that future drives should be able to read it -- a very important factor when considering long-term archiving because who cares if it lasts if you won't have a drive that can read it!
DVD-RAM is great for video editing systems, for which, I bought my DVD-RAM drive to complement my brand new Matrox Marvel G200-TV at the time. Again, much, much cheaper than magnetic disks per MB/GB.
-- Bryan "TheBS" Smith
AT&T's Virtual Network Computing (VNC) is an excellent GPL licesned product to look at. When the server is run on UNIX systems, it allows you to run upto 99 independent sessions (more if you reconfigure/recompile it) on a single UNIX system. These sessions have their own window manager, programs, etc... tailored for the user running it. Essentially, you have a full X-Session running just like if the user was directly in front of the GUI console of the workstation. The user can then connect to the session from any VNC client, whether that client is MacOS, Windows, UNIX, DOS, Palm, etc...
Using VNC instead of a traditional X-Server on the PC side has several advantages. One is the fact that a user can disconnect and reconnect their viewer on another system and not have to logout. This also means that if the VNC viewer system (e.g., Windows) crashes, their entire X-Session is still running (and can be reconnected to). VNC also works fairly well over low-speed connections (as good as X11R6.3 extensions for low-speed connections), provided you minimize background images. An additional advantage is the fact that it runs on a single port (5900 + session #) which makes SSH tunneling extremely simple (side benefit: "low-cost, more secure" remote access than most "all ports open" commercial VPN software).
I work for a company whose applications are 90% UNIX based (Theseus Logic), and that's not likely going to change soon (as EDA tool vendors are choosing Linux over NT because of the true multiuser capabilities). We use Linux and Solaris sytems to run these applications. Although we are starting to dual-boot some of our NT Workstations on our desks with Linux (although my personal workstation and all our servers are 100% Linux ;-), most of our work is done via VNC over to our headless Linux and Solaris systems. With 512MB to 1GB of RAM, we can easily accomodate 10 engineers on each system with fairly intensive engineering applications running. This has additional benefits like accomodating node-locked licenses that normally won't remote display (to another system), but will work in a VNC session (because the VNC session appears simply as display "localhost:vncsession#.0").
VNC is also a great way to slowly move to Linux. Users can spend 6 months becoming familiar with Linux use via a remote VNC session, while still running Windows on their desktop. You can also use VNC to UNIX systems so any user can run those few (or many, like us ;-) UNIX applications when needed. Again, VNC is so simplistically powerful (especially on UNIX systems), you'll never run out of uses for it.
Again, instead of trying to deal with finding a costly terminal server program, or a PC-X-Server, evalute if you really need one. If your UNIX platform supports VNC (and I seriously doubt there is a major or even minor UNIX platform where it has not been ported to), just give your users a full UNIX session with all it's goodies. Best of all, with the VNC server and all its windows/apps running on the UNIX platform itself, you have 100% native execution.
[ Side Note: Unfortunately, since Windows is not a true multiuser system (and only a bastardized one with products like Terminal Server), you cannot use the VNC server in a reversed role (multiple users on a Windows NT system). Although you can use VNC like pcAnywhere (unified single session, remote and local user both see the same and control the keyboard/mouse simultaneously). And like pcAnywhere, VNC sessions can be shared, allowing remote training, etc... (even on UNIX, with multiple sessions running) ]
-- Bryan "TheBS" Smith
I think people realize that Linux is both a consumer platform and a business/server platform. The Windows 98 / Windows NT-2000 platforms are distinctly separate platforms, with each only filling in one of those. Comparing Linux to Windows NT-2000 changes the game radically.
And don't count out forthcoming distros with kernel 2.4.x and XFree 4.x. I think it will level the playing field.
-- Bryan "TheBS" Smith
The argument is clear-cut and to the point. The web was designed to be a platform-independent repository of information and services. The W3C has done an outstanding job to give commercial vendors the tools and standards to even implement extensions to them for vendor-specific documents and capabilities through technologies like XML and DOM. These technologies were designed specifically so vendors could release new technologies and not have to wait on the "slow" standardization process to catch up. For Microsoft not to use these well-defined and documented technologies to implement almost any equivalent Win32 capability is to say they do not believe in a vendor-neutral standardization process, but believe in a Microsoft-generated standardization process.
And that process is the reason why no API lasts more than 2 years in the Windows world, no program is stable for more than 2 days in the Windows world and no independent thought exists in the Windows world more than 2 seconds. And this is coming from a former Win32 developer and enthusiast.
-- Bryan "TheBS" Smith
I trust Borland. They basically pioneered high quality development tools for the PC, while also pioneering the like a book mentality to software licensing.
Understand this isn't Microsoft we're dealing with, but Borland. I trust them to license commercial software and they make high quality software at that. With their powerful Object Pascal language and Delphi system behind a number of quality Windows applications (from largely non-Microsoft partners), it will be nice to get those applications ported to Windows.
And this is coming from a guy who prefers GPL, GNU/FSF, GCC, GTK+ and Gnome itself! ;->>>
-- Bryan "TheBS" Smith
I'm a longtime OS/2 fan and love the Workplace Shell. It's a full-up, truely object-oriented desktop (unlike Microsoft's Active Desktop that only has some O-O-like attributes). I've seen some OS/2 WS-like programs for X, but I'd like to see a full window manager like it. IBM, where's the source? [ wishing too much I guess ;->>> ]
Even without a OSS WS-like environment, can we at least get a file-oriented window manager? Personally, I'd love this. I hate just about everything about Windows with exception that the window manager is file-oriented. I guess that is what the Gnome 2.0 focus is on, but I'd really just like a light-weight window manager that was file manager-based -- even if you had to make some footprint decisions to leave out support for some graphics types, etc... I personally think it's overdue for X.
Anyone agree with me here? I like the CLI (which is *NOT* really a bad thing in UNIX because the shell capabilies are 100x better than Windows, hence why I always install Cygwin/BASH on NT ;-), but that 2% of the time I wish I had a file-oriented window manager.
-- Bryan "TheBS" Smith
I've had issues with older PCI sound cards on newer PCI 2.2-spec mainboards. Nothing new, nothing limited to AMD either.
As far as the FPU bug, it goes in the same basket as the 100+ other errata on any Intel chip. Every mainstream microprocessor has bugs, dozens of them. Heck, my company just released a 8-bit microcontroller and, not even back from the Tawainese fab yet, we've already discovered our first errata (a small one that is easily corrected in firmware/software)! But Intel seems to take the cake and it's not just because they're #1 either.
On with the FPU thing, I've seen some IGNORANT people get on the Quake benchmark BS. I'm sorry, but idSoftware's Quake engine uses integers for a lot of things -- but because the Intel ALU is so freak'n slow (1/3rd the speed of any K6 chip), idSoftware just found that loading integers via the FPU load instruction on the Pentium ended up being faster (even though it was some 4 instructions to 1 do to the same thing -- talk about a design flaw!). As such, even though the K6 would have actually been faster than with the original code, the "Pentium optimized" version results in the crippling the advantages of the K6 in order to better optimize an otherwise poorly designed Pentium.
Remember, "Pentium optimized" means that this is how you must write your software to fix Intel design flaws in the Pentium (at least 25-30% of the time). Intel builds full-custom ICs which means you get the human error factor (although it does have some advantages), compared to many other firms where a lot of design is done by optimizers and other EDA tools.
-- Bryan "TheBS" Smith
This is just like the whole problem Gateway 2000 had with the AMD K6-2 systems -- it is NOT the chip -- but the selection of components around it. With the AMD K6-2 systems, Gateway 2000 choose a sub-standard mainboard (cannot remember the vendor), largely based on cost in volume (and possibly guaranteed volume as well). Now they have what looks like to be a current draw issue that is due to the mainboard or power supply. I mean, did c|Net even bother to check if AMD is using an Athlon-certified power supply???
I personally think AMD's track record as of late is better than Intels! Lest we forget Intel had a mainboard recall on the original SE440BX (reference i440BX chipset) mainboard due to power supplies frying the mainboard and its components. And I'm not going to go into the whole MTH (memory translator hub) fiasco and the RDRAM RIMM signal integrity that preceded it. I really would like to know what mainboard Gateway 2000 is using for this system. If its past K6-2 problem is an indicator, I'd say it's Gateway 2000 going for the lowest bidder.
Which makes me wonder why they pay a premium for their PIII mainboards instead of using someone cheaper than Intel??? Let's see here, for Intel, use costly Intel mainboards instead of cheaper alternatives (unlike others like Micron who uses the more flexible, VIA-powered Tyan S1854 Trinity 400) ... but for AMD, use lowest bidder instead of paying extra for AMD stock ... hmmm ... First the K6-2, now the Thunderbird, that's STRIKE TWO GATEWAY 2000! [ Or was this intentional??? ]
-- Bryan "TheBS" Smith
I think people fail to realize that NASA has been fairly responsible in their space endeavors. For example, NASA only uses decaying radioactive materials in spacecraft that do not orbit Earth, and only when the power requirements of the craft deem it absolutely necessary (i.e. it is impossible to power it with solar panels due to solar distance and weight). NASA even went as so far as to worry about the tiny ammount of radioactives on the Lunar Module of Apollo 13, as to plot a much more risky (to the crew) return trajectory (with additional corrections) so as to not have the LIM crash into anything but a remote and deep watery grave (see "From the Earth to the Moon: Episode 8" to hear more on "the Nuke Problem" as the media called it).
This is in stark contrast to numerous Russian Earth orbitals, including even purely scientific ones! Take note of a Russian scientific satellite that contaminated hundreds of square miles (600 mi^2 I believe) of Canadian national forest when its return trajectory was not accomodated correctly. And I need not mention that it is still up in the air whether or not the Russian military, let alone our own, uses radioactives in Earth orbiting platforms.
Give NASA a break guys!
-- Bryan "TheBS" Smith
I work at a chip design firm (Theseus Logic) and 90% of our applications run on Linux or Solaris all from a single Linux server (although we looking to add a 2nd server or a NetApp box). Outside of those programs used by 30-35 engineers, I spend 75% (or more) of my time messing with stupid Windows applications for the admin staff, a measly 7 people (even though the Engineers are 90% of the traffic and data).
From a $30K accounting package (Deltek Advantage) with its own NT server that cost more than our Linux box (and may require a 2nd one soon for stupid Citrix Winframe), to stupid little $2-5K/each Windows software packages here and their for inventory, stock options, etc..., I'm going up the wall. Especially when updating software (never goes right, unlike our UNIX EDA and other tools) and I pull my hair out. Everytime I bring up ERP I get told that since we've already spent >$50K plus another $50K on consultants, so we're not changing. Of course I brought up the point before we spent this money so I get the underlying "crying over spilt milk" or "quit rubbing it in my face" attitudes nowdays.
I can argue TOC with ERP, but for companies like mine that have already spent >$100K on disseparate Windows packages and don't want to pay anymore, a free/OSS package is the only way to get it in house. I sure wish companies would realize that maintaining disseperate little (and even big) Windows programs are just a pain in the @$$. I'm sorry but all it takes is 6 months of UNIX administration and sysadmins realize that UNIX maintainance is just 10x easier (thank God 90% of our engineering apps run on UNIX).
Thank God projects like GNU Enterprise and the Java-based Kontor Project have sprung up. I'd say if you want to help Linux get inside corporate America, look to donating your time on these projects. And you don't have to even be a developer to do so, I'm sure both projects are looking for a lot of bookeepers and accountants for most of the design.
-- Bryan "TheBS" Smith
You no longer have to buy a proprietary x86 OEM solution to get massive, multiple memory and PCI busses for high I/O throughput. The new ServerWorks (formerly Reliant Computer Corporation) ServerSet III-series chipsets are making their way onto retail mainboards from major vendors. They put the disk controller on a different PCI bus than the Ethernet controller.
For those interested, here is a comparison of various chipsets and their aggregate memory and PCI throughput in MBps, respectively (sorry about the format but ./ doesn't seem to allow tables or the pre/code tag):
Again, this is just aggregate throughputs. RDRAM is faster at bursts, but SDRAM has lower latency (i.e. better), and the old EDO of the i450NX chipset doesn't even get close to what SDRAM and RDRAM can do.
But note the massive PCI throughput of the ServerSet chipsets from ServerWorks -- due to their 3 independent PCI busses, of which, 2 are 64-bit (both are 66MHz capable in the HE/WS). They're so good, that even Intel is adopting the 4-way IIIHE for a forthcoming SDRAM server mainboard instead of RDRAM. This is largely because the 4 RIMM slot i840 can only support upto 2GB of RDRAM whereas the 16 DIMM IIIHE can support 16GB. As of right now, all Intel can offer for servers is the 3-year old i450NX so most tier-1 vendors have opted to work with ServerWorks instead.
-- Bryan "TheBS" Smith
The ~4 phases of software engineering:
First off, it's really up to the programmers whether or not formalize their organization and follow the two and even third rigorously. I think he has made an analysis that all Open Source development is done in an unorganized, non-traditional software engineering environment and all commercial development is.
But let's say for a moment his assumption is correct. This brings me to the last two phases.
In stage three, commercial interests can drastically cut short the testing phase. The "rush-to-market" attitude prevails as software quality takes a back seat to profit margins -- the great conflict-of-interest that greatly reduces consumer benefit of the product. Vendor's release early way to often. A common misconception is the fact that while the "release" of an OSS product is less definable, because the source code has always been available, if software bugs are inherit to programs, why not always always make it available then? I think he fails to realize the testing _and_ recoding are inherit in use of an OSS product. And OSS projects still have to have the traditional "feature freeze" milestone to move even close to production quality. Unfortunately, commercial software cannot be recoded by consumers once released which makes it actually less tested than OSS -- again, especially if it falls victim to the "rush-to-market" mentality of most commercial software firms.
Then there's the final stage, which is 85% of the traditional software engineering model. Unfortunately, in today's shrink wrapped software world, it has been reduced to 5%. Why? Because there is basically no revenue stream in fixing commercial, shrink-wrapped software. As such, it does largely go unfixed, only being bothered with if a large volume customer complains enough and is paying several millions or tens of millions of dollars in support costs. OSS support never ceases because there is no set "release" of code, it's always been released. As such, OSS is in a continual state of support.
And that final stage is the biggest point. Again, the traditional software engineering model has always been about 85% post-release support. Only those design teams, whether they are OSS or commercial, who are driven by post-release support (either ethically or even financially) can accomodate this most important phase best. That's why when RedHat and other vendor's say "service-focused software firms" are the key to good software products, they are hitting the mark!
The days of rushing the 3rd phase and ignoring the 4th phase of the traditional software engineering model are over. OSS is here and it's our best chance at releasing quality software because it's inherit design methodology accomodates all phases of the software engineering model, even if it lacks an "official source code release" and the code can be chaotic at various points.
But even the concept and problem of the "choatic code state" is debatable. I mean, do you trust software just because it has been "officially released"? Or do you at least hear some peer-review before using it? I think the later applies to the adoption of all software more than the former.
-- Bryan "TheBS" Smith
Our Sunnyvale office tried to buy some Dell systems with Linux pre-installed. They kept getting the same answer over and over, Linux only supports 128MB of RAM (not the model, but Linux!). They could not even get the memory part number from the sales department so they could call the parts department to order more. It was ludicrious!
So our SV office made the mistake of ordering another model with Windows NT and ran into mega problems with everything from the video card to those sorry-@$$ Promise cards. Sure enough, someone finally contacted me (the order happened without my knowledge) and I tried to straighten out the whole fiasco (still working on it).
At our main office (where I work), we've dropped Dell completely because they have their head up their @$$ for the last 18 months -- seriously. I know they've had a lot of sales, which means they are overloaded and overworked in most departments, but Dell's recent expansion has left them in a cluster-f--- state, unlike the massive expansion of 1994 that Gateway 2000 underwent that was at least 3x as bad, but GW2K still kept good records and kept the order. After a year of Dell shipping systems to the wrong address (although billing to the right one, we couldn't even get it f---ing changed to be the same as the billing after 2 written letters!!!), shipping us replacements weeks late, and then shipping 2-3 of them (which only drives up their own costs!) and, lastly, because treated us like a f---ing 2-year old when our CEO's hard drive went belly up, with 3 different departments all pointing fingers at each other and no one to talk to, we dropped them. I mean, when you give my superiors some lame @$$ excuse time after time again, telling them flat out you don't know anything and can't do crap even though you purchase 6-figures of product a year, you drop a company like that. We gave Dell at least 2 dozen "chances" and they failed every single one ... continuously.
I think everyone knows the moral of the story here: freak the direct PC OEMs. Stick with either established Linux vendors like VA or use some "real" enterprise PC OEMs like IBM, HP and the like. They don't pop hard drives in like bread in a toaster, they'll customize your solution.
-- Bryan "TheBS" Smith
For some reason, StarDivision (and now Sun) continue to assume that only one user will run StarOffice. While I have to put up with that kind of crap in the Windows world (e.g., most Windows software is ignorant of the multi-user issues of Windows NT/2000, including most of Microsoft's own software), you'd figure Sun would know better. As such, when I first installed StarOffice 4.0 a long time ago, I installed it as root and always had issues afterwards when running as other users (even when I opened up the permissions).
Of course, the multiuser installation is hidden with the command-line option "/net". Why this is either not the default or presented in the GUI as an option (at least they _do_ put it in the GUI readme now), I have no idea. When you install with the "/net" option, you install a shared copy that is just as good for multiple users on one system than multiple users over a network of systems. In fact, the "/net" label is really a misnomer, it should be "/multiuser" or something similiar.
The only other thing to remember is to install in another location than "/root/OfficeX" -- something that Sun should change when the "/net" option has been specified at the command line. While it is recommended that after this "multiuser install" each user run the ./setup binary that is located in the shared StarOffice binary directory, it's completely unnecessary. Due to the power of UNIX's multiuser orientation, StarOffice (at least 5.x does) notices that you haven't done the ~2MB user-installation yet if you run the ./soffice binary in the shared StarOffice directory. Simple no? [ Why not the default then?!?!?! ]
The last complaint is the continued political BS regarding the bundling of Java with StarOffice. If you have Windows, it is bundled. If you have Linux, it is not. [ Hmmm, have to check out Solaris and see if it is. ] That is just utter BS and seems to be a political issue with the Java license -- you can bundle it with a Windows binary product, but not a Linux binary product. Nuts!
Now on the bright side, one very cool thing about the new StarOffice 5.2 installer I just noticed is the lack of the requirement of the DISPLAY variable to be set correctly. E.g., I just su'ed in a terminal to root to install via the "/net" option and StarOffice figured out that it was running on the local system and set the DISPLAY variable correctly. Sure wish other vendor's installers would do the same. [ Now what happens if I try the "/net" install without being root? ]
-- Bryan "TheBS" Smith
Not only is there the standard waste in packaging, but what about the waste is shipping costs, store and shelf space before sale and the like? This level of unprecedented waste goes well beyond the "killing of trees".
You'd figure that resellers with a physical store front would have been complaining for years, let alone today's software warehouses. I mean, by reducing at least the 3rd dimension, depth, you can package 3-8x as many units while still preserving the marketing value of the 2-D frontal image! You don't have to be a rocket scientist to figure that one out! Dooh!
-- Bryan "TheBS" Smith
If you're like me, you restribute the CDs on a quality CD-R (e.g., I like Kodak replicator grade) with a nice label, ask $5 for it, and give out 2 freebies to new users for that 1 person that pays you. I only "ask" $5, but do not require it. As such, would that not be a "donation" and separate from the sale?
So I see it as this: I'm merely taking a donation and/or just trying to cover my costs (of which, I actually lose because at ~$1.66/each, the cost of the CD-R, label, color ink, printer, PC, etc... ends up being more). Again, I don't require people to pay for it either, but most people that are asking for their 2nd, 3rd+ CD feel they've already had their "free one" and spot my $5.
I also consider the distribution to be "local", kinda like if you were making copies for and to yourself. I just do it in my local LUG(s).
Lastly, I do download the source, just in case anyone wants it. It's available upon request. And maybe that is how you should approach it, "available upon request." Hmmm, need to read the GPL on that one.
-- Bryan "TheBS" Smith
While I'm not totally going to dismiss this guy's viewpoint, he should really try to explaining further some of his Linux "assumptions". They are "assumptions" because I they do little but prey upon the common myths about Linux. E.g., such as the fact that without a single GUI framework, no one likes it.
I think MacOS X will be a great addition to the OS landscape, especially if Apple keeps the codebase as close to FreeBSD as possible. But I don't see everyone going out and rushing to buy MacOS X unless it gathers some more and heavier hardware support on its own, single platform.
-- Bryan "TheBS" Smith
While the film had technical holes all over it, I found the underlying interactions between a conquering species and man to be interesting and consistent.
For example, the viewpoint of the aliens of human conversation, and vice-versa. The general distain and disgust, ignorance and indifference. Nothing breakingly original, but I noted on many occassions that the dialog and actions varied from the traditional interactions of two people (or species in this case) that could care less about each other. Again, I found it very consistent in many point, with the stupid technical issues interrupting it (to the point of my disgust!).
But what really drove me was all those technical impossibilities. Like the Harriers, bombs and computers that lasted 1000 years. The "learning machine" that thought the man to speak their alien lanaguage, but the aliens did bother to learn man's. Although that could be more of the interaction element I was talking about -- but still a stretch in application.
All-in-all, the movie would still have to be radically re-written to make it believable. Even if this meant that mankind "loses" -- it still would have been much, much better than with the general Hollywood ignorance and ending.
My $0.02 ...
-- Bryan "TheBS" Smith
I concur with just about everyone here that the Linux Router Project (LRP) is a floppy solution that can run on even a lowly 386 CPU. You should be able to find such a system for $50, and not have to spend the $$$ you mentioned.
Otherwise, if you really don't want to use a PC, I'd grab something like the SonicWall SOHO/10 for around $400. As of last year, SonicWall's products were the only ICSA-certified firewalling products for under $4K. The SOHO/10 is a little 25MHz 68300-powered Coldfire running some RTOS (probably VxWorks). The SOHO/10 allows upto 10 nodes transparent access out, and even provides one-to-one NAT (private-to-public IP mapping) if you want to share out services, which you can filter, of course, by service.
Just FYI, their high-end product, the SonicWall PRO, is powered by a 233MHz SA 110 StrongArm chip and features a myrid of VPN and encryption options built-in, along with a DMZ port. It lists for $2995, not bad for its capabilities. But I figure you're not looking to spend THAT much. ;->>>
-- Bryan "TheBS" Smith
Dude, I think most MacOS and Linux users and the applications for each systems are quite different.
To take your example, I would tell the home user to get a system with Corel Linux PRE-INSTALLED instead of trying to install it him or herself! And there are other issues (again, I didn't read the whole thing because it made me sick).
There was such a narrow focus (like the fact that you can get Corel pre-installed) and a lot was left out that it was just pitiful. Add to the fact that most home users would be better served by Windows instead just kills me.
Again, most MacOS and Linux users are quite different beasts, at least at this point. They should have never been compared as there is little overlap in their markets.
And I do not consider "anything but Windows" to be a good argument or market. Apparently this is what the media keeps harping on, and that makes Windows advocates hate Linux even more.
And yes, I use Linux everywhere, at home and work because I'm not even 25% as productive in Windows.
-- Bryan "TheBS" Smith
Guys, I'm sorry. I have never seen a more poorly written article on a comparison that should have never been made -- just exactly who was this written for?!?!?! I don't know what c|Net was thinking ... maybe "Hey, let's compare the two non-Windows OSes that we think are easiest," but they even failed that.
I mean, the very first comparison, "installation," that was just rediculous!!! Could they at least find a PC vendor that pre-installs Corel? If not, then it is NOT a good comparison!!! Duh!
And who "wins" a category? Com'mon, you didn't even define how you choose. I mean, how did you weight hardware compatibility? If MacOS was written for Mac hardware (which means just about everything is supported), how did Corel win if it doesn't support all Intel hardware? What do you mean it's "less expensive"? Now you are jumping criteria! Please define them beforehand! So many laws of comparison were broken here.
And that's what this eventually boiled down to. Too many variables against extremely different platforms. Geez, just use the system that runs your apps.
I'm still trying to figure out who this article is written for?!?!?!
-- Bryan "TheBS" Smith