Fog Creek Software
Discussion Board

Windows' Biggest Enemy Is NT, Not Linux

Some rag jocks are finally waking up to what the real world (including Microsoft) has been saying for years.;jsessionid=SRGJLOGH4VRKSQSNDBGCKHSCJUMEKJVN?articleID=7900043
Microsoft's biggest enemy is its own shadow. If they think NT4 was a good contender, Win2k must be almost invincible. Look at it this way: here you are sitting there with a few 1000 Win2K licences that are valid indefinetly. They have been paid for and are written off long time ago. You have all the inhouse knowledge for developing for this platform. It is in place and performing well. Thanks to the likes of Intel, Western Digital, ATI and a cohort of others that same Win2K licence is even doubling in performance every single year.
So it is free (already paid for), we can do everything we want with it and it gets faster and faster everytime we replace a box. What is going to beat this?

Just me (Sir to you)
Wednesday, March 19, 2003


Same goes for Office 2000.  MS is letting almost the entire world qualify for the less expensive Office XP academic licenses just to move some boxes.

Note also that Office 2000 and Windows 2000 are the last in each series that don't "phone home" to authenticate after installation.

Mitch & Murray (from downtown)
Wednesday, March 19, 2003

Threats change; so should you.  NT just can't be changed as dramatically as the latest Microsoft OS, nor is it as forward-thinking.  The .Net CLR is in a way like a complete security rewrite of Windows.  Which one has a greater chance of weathering the unknown threats of the future?  Definitely not NT...

For a more quantifiable comparison of Microsoft OS's from a security standpoint, check out the relative attack surfaces of their OS's 2/3rds of the way down this page:

If it were only about cost per non-security feature then yes, Microsoft may have a problem.  But the cost of potential damage in the face of constantly changing threats means they'll probably do just fine, thanks to the greater vulnerability of their older OS's.

Wednesday, March 19, 2003

I agree that security updates is about the only compelling reason to upgrade. Many of these issues (such as the atttack surface etc.) can however be corrected by good management processess and some extra (less expensive since it only has to concentrate on this one issue) software.
The big problem is that the current platform has become more than good enough for 99% of the businessess out there, so there is no reason to stay in the upgrade cycle.
OS'less PC's, legacy "unlimited time" licences and high quality of past offerings are the number one enemy of Microsoft.
The idea of a business in a situation were they already own a good system switching voluntarily to another platform becomes almost laughable, and the real growth seems out of the market (at least in Europe/US). That is why MS has to spend billions on research to find new killer apps. They have to reinvent new growth markets, since they killed the office software market if they can't get it into a subscription only model.

Just me (Sir to you)
Wednesday, March 19, 2003

Incidentally, there are also many good reasons *not* to upgrade. 

The phoning home is one, but not the most important - it's the bugginess of the XP family that is the real killer for us(especially in Explorer and Word).

For instance, moving files in Explorer, if you move your mouse off of the folder to which you are moving it (sometimes inevitable if you are rightclicking and selecting from the menu) - the file ends up going to whichever random folder the pointer is over. (Prompting a copy-only policy after one too many file was "lost").

If you are accessing files over the Internet, the directory listing starts to refresh like a maniac (once every 2-3 seconds), meaning that you can't click on the directory or do anything.

Or how about the fact that Win XP can no longer do WebDAV if authentication is enabled?

And don't talk to me about random unexpected behaviours and crashes, both of which are a much greater problem in XP than Win2K for some reason.  I don't know if it's just more in your face ("can we send an error report to Microsoft") or what.

I'm not sure that previous "good enough" releases are really the problem as much as the newest release(s) have apparently been slammed out the door too soon.  The release cycles have also sped up in the last few years/decade, meaning that the changes between each new version are not as dramatic.  Coupled with the need to show continued and increased "growth", Microsoft is in an unenviable position - and I'm not sure it's one that leads to better quality software.

Wednesday, March 19, 2003

I'm still on Win2k and see no reason to upgrade to XP or Whistler. IMHO Win2k is "clean" and stable, and does everything I need.
Recently a manager at a client I sysadmin for was rumbling about the cost of upgrading to Windows.Net and Exchange 2003. I wrote up a two page point paper that boiled down to "who said we're going to upgrade, and why would we want to?"


Wednesday, March 19, 2003

I disagree that there's no reason to upgrade to XP. There's one HUGE reason for LCD users: ClearType. As a guy with 5 LCD monitors (2 on one PC for me, 2 on one PC for my wife, and 1 laptop), that is reason enough to pay.

Plus, there's ongoing hardware issues. As OSes become "obsoleted", new categories of hardware devices become only supported (or only supported _well_) on new OSes. W2K is still current, so it's not really behind the curve here, but it will be eventually. W2K added things like full USB and UDF support that were missing from NT4.

I suppose if your computing world stands still (same PC, no new devices), then your OS can, too. That's not likely to be the case, even with the glacial upgrade speeds of corporations (who will typically lease for 2 or 3 years, giving a 2-3 year turn-around cycle for PCs).

Brad (
Wednesday, March 19, 2003

While from a home perspective I appreciate a lot of the changes in XP (relating primarily to digital video and imaging), and the startup time is very impressive, I will say with absolute conviction that I have not seen a _single_ user interface change in XP that is for the better, but instead it all seems to be change for the sake of change. From permissions (what the hell is with the "type in a name without a lookup and we'll verify it for you" BS?) to the control panel, it's all just a giant leap backwards. I say this as someone who embraces change when it's an improvement, not as a crusty "who moved my cheese?" user, but if one could bring the kernel/subsystem changes from XP to 2000 and can the UI (yes you can go to "classic" mode, however subtle irritants persist), I would jump no that in a second.

Jimmy Chonga
Wednesday, March 19, 2003

Oh yes, the hardware driver issue. Well, if corporates keep running W2K until 2020, will this not be a huge incentive for any hardware vendor that targets that market to keep fully compatible with W2K?

Just me (Sir to you)
Wednesday, March 19, 2003

I run both XP (laptop and work machne)and 2000 (home desktop) and have a spare laptop that runs on W98SE

99% of the time I don't notice which OS I'm using. The 1% I do I must admit I prefer XP slightly, but certainly not enough to go to the trouble of upgradiing,  even upgrade if it were free.

XP does appear to be a lot more buggy, though it does recover from the application crashes, whilst the only time W2000 crashed on me I had to restore the Ghost Image.

I like the fast boot up time but has anybody else found that you are often not able to do much for the first thirty seconds or so after the boot. I suspect that there are plenty of things loading that used to load before the desktop came up.

Stephen Jones
Wednesday, March 19, 2003

I feel that Win2k has finally gotten to a point where it is stable so I don't want to change. Unfortuantely they will eventually phase out support for it. But I will run it on the last service pack as far as it can go.

I am looking forward to seeing .Net server in action (or whatever they call it this week).

Ian Stallings
Wednesday, March 19, 2003

Typical MS behaviour.  The simple fact is, their past products are their biggest enemy because they haven't done much of any significance.  WinXP offers very little over Win2k.  Certainly, in the performance category, it is just more bloat, more gloat, less real performance.  The simple fact is, every new version of MS Windows has gotten more bloated than the last, thus performing worse than the last.  Win98 was barely an improvement over Win95 -- perhaps a little bit faster, and no real significant UI changes.  WinME was a downright slap in the face to consumers by MS:  not only does WinME offer nothing over Win98, but it even loses a few things, like real-mode DOS (the only advantage of WinME I've been able to discern is faster boot-time, due to not having real-mode DOS).

Then they complain that people aren't upgrading, so they come up with these lame contrited tactics to try to force people to upgrade, such as phasing out support for older products (such as Win2k), which are essentially the same support-wise as newer ones.  I suppose I shouldn't be singling MS out, since all software companies try to do this.  Adobe's another disgusting example.  For the average user, what does acrobat 5.0 offer over acrobat 4.0?  Nothing.  It just takes a lot longer to load.  What about Word2k over Word98?  Again, nothing of any significance.  Likewise with WordXP.  No, take that back.  They have added the particularly annoying M&M "character" to help you out.  Another slap in the face to the consumer.  Apparently, in MS' mind, consumers are too dumb to go to the HELP menu, so they'll give them a little cartoon, as if they were children.

It's a clear and obvious sign that MS has simply made Windows more bloated that companies wont upgrade to WinXP because of hardware-concerns.  If WinXP was really "better" than Win2k, it would run faster, take up less memory, and be more stable on the same hardware.  Now, I know some are going to say, "by that logic, the old CLI interfaces are better than modern GUIs".  Well, for people who only need CLI-interface functionality and the accompanying software, yes, it is.  But that's a major exageration of the issue.  We're comparing Win2k/NT and WinXP here.  What, exactly, does WinXP offer that I need or want, which would justify me paying lots of money to upgrade my hardware, or just accepting inferior performance on the same hardware?  I'm just a normal home-user, so I can only answer that from my own POV:  the answer is nothing, which is why Win98 is still on my windows partition, not XP.

This, of course, is part of the reason why I like FS and OSS software development.  They don't feel compelled to add extra "features" which I don't need just to compel me to upgrade.  First concern is usually with fixing bugs, stability, performance, etc.  And usually features are only added when they actually provide something that would really help most users get their work done faster.

From a technical standpoint, there is ONLY one criteria to consider when asking "what is the best program" for a particular application type:  how fast it allows the user to get the user's task done.  So what if MS Word 2k offers a fea new features over (say) Word 95.  The vast majority of users don't use those features, for one thing.  For another, MS is adding "features" which actually hinder the user, by fighting his/her decisions.  Also, the additional bloat makes programs load slower and take up more RAM, which will be an issue when the user is using several RAM-hungry programs at once.  All of this translates into the user not getting his/her task done as quickly, which means less productivity, which means an unhappy user. 

Oh yea, the enormous amount of security holes that come with new MS products is also a non-plus.  Security holes means the possibility for viruses or trojans to destroy your data.  That means you have to spend valuable time dealing with the problem:  purging the virus, restoring destroyed data, re-installing any programs which may have been damaged, or dealing with other nuisances.  All of this means that the user is less productive, and gets less done on all of the tasks (s)he was doing.

Summarily, MS had better have some rock-solid reasons for me to upgrade to their newest products, and these reasons better start out with me getting my work done faster and being more productive -- all things considered, including load-time, memory-consumption, security, stability, bugs, and other performance issues, not just the user interface and new features (most of which I probably won't use).  Unless MS can show me how upgrading to their latest releases is going to help me get my work done faster overall and be more productive overall, they can go fuck themselves.  They can also go fuck themselves if they have the nerve to ask me to upgrade my hardware to run their new bloated software.

Wednesday, March 19, 2003

potty mouth...

Wednesday, March 19, 2003

I wrote up a two page point paper that boiled down to "who said we're going to upgrade, and why would we want to?"

Philo, can you mail that paper to me, would be interested in the why part.

Prakash S
Wednesday, March 19, 2003


It's pretty silly to complain about the extra unwanted features in Windows or in MS Office and then turn around and promote OSS software as an alternative. If you take that point of view, your average linux distro is guilty of the most awful kind of bloat.

I'm currently running RedHat 8.0, and, as part of the standard distro, I have 2392 files in my /usr/bin directory. There are another 87 files in my /bin directory and 165 files in /usr/bin/X11. And yet another 165 files in my /usr/X11R6/bin directory.

Most of the time, I only use 10 or 20 different linux commands, and I know all of these off the top of my head. But sometimes, when I need to do something weird, I have to use apropos to find the right tool, and I have to search through the obscure descriptions of dozens or hundreds of tools. That doesn't make my work faster.

Anyone who complains about the 80/20 problem in Microsoft Office and turns to linux solutions as an answer is missing the point.

So is the proliferation of unix tools an example of software bloat? Why would such an insane collection of tools grow up in an environment that's purportedly unmotivated by profit, if profit-seeking is the source of all software problems?

I don't know. You tell me.

Benji Smith
Wednesday, March 19, 2003

Microsoft is very much aware of this problem. Heck, they've had to extend NT's end-of-life date a couple of times already. They finally came to understand that users will not upgrade just for the sake of running latest and greatest, but instead need a compelling business reason.

For MS Exchange, the penetration rate of Exchange 2000 is less than 50% of servers after 2+ years. In most cases, people running 5.5 didn't see a need to upgrade to 2000 and deal with Active Directory headaches. So for the next release of Exchange MS is putting in feature that make a huge difference to admis in terms of server requirements, managebility and user experience.

I'm sure that next release of windows (probably after .NET server) will take similar approach to make sure people have real reasons to upgrade.

Wednesday, March 19, 2003

Actually the same bloat and feature creep tend to happen in OSS too - have you seen Mozilla or emacs lately?

Wednesday, March 19, 2003

Is it impossible to get ClearType except when MS bundles it with an OS?

Li-fan Chen
Wednesday, March 19, 2003

Since I've never used RedHat, I can't really comment on it.  However, if it has an option for "custom install", then your entire bloat argument falls apart, because those thousands of files you complain about could be removed by choosing not to install them.  The hundreds of commands you don't use, you can delete (if that is, no other program reliese on them).  In regards to invidual program bloat, some programs are adding unnecessary features (e.g., KDE).  Others, however, such as and Dillo, are sticking to the important features, and not implementing trivial rarely-used ones.

You cannot claim that the proliferation of apps and software tools if "bloat".  Bloat refers to a single application, one person or group's responsibility.  Many people are using GNU/Linux, and some people have obscure needs, while others are unhappy with the standard app.  Naturally, there's going to be a proliferation of apps.  RedHat may be responsible for the bloat you claim exists in its distribution, but only if they give you no other choice.  If you have the option to "install this, or not install that," then it hardly falls on RedHat, but on you.  But to answer your question -- how can such bloat proliferate by a group unmotivated by profit -- the answer is that your question is false ;-).  RedHat is a corporation, motivated by (at least partially) profit.  It should also be noted that most distribution (don't know about RedHat and the other one's targetted for Joe-consumer), but one's like Debian, Slackware, Gentoo, and other purist distributions, give the user complete choice over what's installed and what's not, and can still be run on a 386, if the users needs are that minimal.

Also, I wasn't complaining about MS adding extra useless features (most of the features can be chosen not to be installed anyways).  I was complaining about them having the gall to add extra useless, rarely used features and create a product with inferior performance (load-time, memory-consumption, etc) and ask me to pay them for it.  No, try to force me -- by phasing out support on older products -- to pay for it.  That's what I'm upset about.  In the FS/OSS world, it's very rare in practicality that you have to pay for upgrades (though it is certainly possible and allowed by the GPL, see Real-Time GNU/Linux).

Let's take MS' real flag-ship money-maker product:  Office.  I have used Office 95, Office 98, and (currently), Office 2k, which came with my new computer.  Let's just skip 98, and compare 95 with 2k.  Office 2k literally offers nothing for me over Office 95.  I do not get things done faster in 2k, but in fact slower, because 2k insists on fighting me (if you don't know about how Office 2k fights what the user wants, you haven't been using it long enough).  So, how, exactly, is it that MS has the nerve to ask me to pay extra for Office 2k, when it's offering nothing for me, but actually hindering me?  Now, just for document-creation in GNU/Linux, let's look at my alternative:  LyX.  It does not fight me, and allows me to focus on content rather than formatting.  The result?  Professional-quality documents that are properly formatted given their document type, without me having a headache.  They even use rational intentations -- 1.5 to 2 inches -- so that documents are easier to read.  1" margins all the way around may "look" nice, but documents aren't meant to be hung on walls; they're meant to be read, and it is alot easier to read a short-lined than a long-lined document).  To prove my point, look at this web-page.  Notice how the text doesn't scale to fit the entire browser, no matter how large you make the browser?  That's for readability.

Summarily, what bothers me is that MS has the nerve to ask me for my money to upgrade to what is an inferior product for my purposes.  Not only do they have the gall to ask me to do that, but they have the arrogance to try to force me to do that, and everyone else as well, by preventing OEM's from allowing consumers the choice between Office XP and any other version on their computers, and by phasing out support for older versions.

Won't work for me, because I have no need to ever upgrade my windows software.  I basically use it only for entertainment (read games), and I won't be buying any games in the future aside from Tomb Raider 6.  I also have no use for technical support, as I almost invariably know more than the tech-support guys I call, who's solutions to problems are obviously read from a cook-book, and ultimately boil down to "reinstall everything" if all their little cook-book steps fail.

Wednesday, March 19, 2003

Perhaps, if you think Office 2000 is making it more difficult for you to work, you should RTFM. Features like AutoCorrect can be customized to a very fine-grained degree, and you can make your work go much faster.

Benji Smith
Wednesday, March 19, 2003

Don't you remember what Joel said?  User's don't -- no, can't -- read the manual.  Especially power users.  Especially in GUI environmnents.  I quickly figured out how to customize auto-correct to my liking..  There are many other very annoying irritating "features" of Word which fight the user, many of which CAN NOT be turned off.  The most annoying one for me is whenever you insert an image, and try to place it somwhere.  Often times Word will fight you and put it elsewhere.  When you finally have it where you want it, it can be screwed up by every little enter you press. 

Oh yea, if you group objects together and move-them, sometimes word will mess up the relative distances between them.  My favorite is (on a Mac), when I tried to insert a phylogeny into word from a Canvas file.  If you know nothing of phylogenies, they usually employ perfectly straight lines (0, 90, 180, or 270 degrees).  I past the file into word, and it makes some of the lines at slightly off angles, resulting in a very unprofessional phylogeny:  crap which is unpublishable.

Wednesday, March 19, 2003

I disagree that XP provides no UI improvements.

I prefer the start menu UI on XP over Win2K.  There are little improvements like having the start button go all the way to the corner -- this removes a little irritation with previous Start menus where clicking in the extreme corner of the screen wouldn't open the menu.  Anyone who's read "Tog on Interface" will see this as an advantage.  I also like the frequently used program list in the start menu and the ease at which you can "peg" apps to stay in this list.

I like the drab green Luna look better than the mid-90's 3D look of Win2K.  I like the integrated 802.11b roaming support.      I like the way Explorer has integrated new columns to let you manage multimedia files -- it can read MP3 ID3 tags and WMA attributes and sort files by title or album, or it can show version and copyright info from app/dll resources when you're looking through the system folders.

XP isn't perfect... the whole hiding items thing it does is annoying, but can be turned off, and some of the changes are pretty minor.  If I was upgrading business systems, I'd probably not think it was worthwhile, but on my personal machines and new equipment, I find it worthwhile.

Ben Combee
Wednesday, March 19, 2003

With a few exeptions most of the OSS software does not have the concept of separating the "new features" from the "patches". Yes, there are a few "profesional" OSS shops that maintain old branches over some time, but not unlimited as you imply, and probably not longer than a typical MS support lifetime.

Just me (Sir to you)
Thursday, March 20, 2003

Dear dh003i,
                    While you are right to say that the incremental changes MS has been making in its software since 1995 often are not reason enough for people to go to the hassle and expense of an upgrade, most of your detalied examples are contentious, or just plain wrong.

                      First let's cut out the MS/Linux flame war. You can't complain a power user doesn't RTFM in Word and expect him to go through all the /man files in Linux. A full scale Linux distro is huge; it does has the advantage of being customizable but that is a double-edged sword.

                      There are in fact considerable changes from W95 A to Win 98 SE (Win ME was a  lemon, although Windows file protection was a signiificant new feature). Most of the changes were behind the hood, in increasing stability, more stuff becoming 32 bit, and better utilities. True MS didn't market them but marketed all the irritating features instead . I always advided people that if they had W95C there was no reason to upgrade, but if they had the earlier versions then they should do so when they did their periodic format and uninstall.

                      Now of course, with W2000 the yearly or six-monthly reinstall is a thing of the past; in fact you probably only boot the machine that often (which considering how long it takes is probably a good thing). This is the problem MS are facing on the OS front. They have produced something nearly everybody likes (if they hadn't we wouldn't have bought it) but are now finding that people don't want to push their luck and risk something else, particulary when XP is basicially NT5b. Originally W2000 was going to replace both NT and W9*, hence the name, but MS couldn't get the backward compatibility with W9* ready in time. They issued ME (W98 + new name) as a stopgap, and then reissued W2000 as WinXP when they'd sorted out the problems.

                  With regard to Office 95 compared to Office 2000 you are plain wrong. First of all there is the question of Unicode support, which is in itself reason to buy it for half the world's population, though obvioulsy at present a much smaller proportion of the world's computer users. With W97 on Win 98 I would regulary find that 10% of my Word documents would not open on Arab enabled windows if written on non Arabic enabled Word, and vice-versa- remember I am talking about documents produced in English on both programs. Secondly Word 2000 bought iin an HTML converter that worked (even if it seemed to do so by producing a style sheet with more entries than the book of Leviticus); if you needed to convert your Word resume in HTML on the fly tnen it was great. Table support was also better.

            Now in general MS revamps one of its programs for an ugrade (I believe the main work in 2000 was done on Access), leaving much of the rest of the Office suite with cosmetic changes only, but one thing that is happening is increasingly smooth automation between apps. If I need to print out an attendance sheet or other tabular data I keep in Access I simply convert it to Excel because for printing tables easily Excel , well it excels.

              I do agree with you that the default settings for Word can be incredibly irritating, and one of the main problems people ask me is why Word keeps on doing what it does even if they haven't told it to. I find out that tabbling and entering numbered lists manually to be much quicker than wasting hours trying to work out what combination of keys I need to press to keep the list numbered but to have a second paragraph wihin the same number indented as it should be. I suspect that some of this is a case of MS listening to what the customer says it wants ("Carol really has no idea about punctuation - can you get the program to put in the capital letters for her?"). MS does carry out fairly thorough testing on these kind of things, so it can't piss off everybody (just nearly everybody!).

Stephen Jones
Thursday, March 20, 2003

Dear Stephen Jones,

Your point are interesting, so let me look at them one-by-one.

In regards to RTFMing, I was referring to GUI environments.  Because an environment is a GUI, everyone's first assumption is that you don't have to RTFM.  People who write GUI programs should assume that (on the other hand, people who write CLI programs should assume just the opposite, and build in good manpages).  There are some GUI programs which you might actually need to read the manual for, but I haven't encountered many of those outside of the scientific area (e.g., definately have to read the manual for ClustalX, if you want to generate good alignments).  For most GUI programs, you should be able to figure out how to do something just by looking around.  If you can't, something's wrong with the program.

Also, let's draw a distinction between a criticism and scorn.  I will criticize any FS / OSS program which isn't easy to use, doesn't have good defaults, not enough customizeability, or too much bloat, and make constructive suggestions accordingly to the authors.  However, since I am getting the program for free -- usually as in freedom and as in price -- I have no grounds on which to be contemptuous.  Someone gives you a gift, you don't say "well, that's not good enough for me".  As for proprietary providers, hwoever, the story is different beause their programs are neither free as in price nor free as in freedom.

So, I'll give constructive criticism to any FS / OSS program-author on things which are wrong with the program and ways to improve them.  But I won't be contemptuous if they put out an upgrade on the web which doesn't offer much over the old version.  However, if MS is going to ask me to pay money for a new program -- worse, if they're going to try to force me to buy their new programs through various tactics -- I am very well going to scorn them if these new products don't offer anything of significance for me.

Of course, I'm analyzing all of this from my own point of view, which is really the only pov I can concretely analyze anything from.  For reference, I place myself as a home-user who's needs include web-browsing, e-mail, file-sharing, music, movies, document-creation, spreadsheet/graph-creation, database-creation, newsgroup-reading, FTP-uploading, gaming, and various biological scientific needs.  I all of those needs, except for gaming, using GNU/Linux; WindowsME, for me, is for games.  I also have the typical productivity applications on my Windows partition.

So, let's address the first thing.  Why did I even need WinME?  The answer is, I don't.  It provides nothing over Win98, except for booting faster because it doesn't have real-mode DOS, which is a disadvantage.  But, WinME came with my computer and I don't feel like uninstalling it and re-installing Win98 just because it's a little bit harder to get TombRaider1 and Prince of Persia 1 & 2 to work.  So, as far as my current needs, WinME basically is fine for them.  Why, exactly, would I upgrade to 2k or XP?  2k is a stable environment, though still frought with MS security holes, but I use GNU/Linux for all of my productivity, so why upgrade to 2k when it's likely all of my games won't work?  There' also XP, under which many, but not all, of my games will work.  Again, why upgrade, when many of the things won't work?  Why shell out money (what is it, $100 bucks for a home edition?) for XP, when ME suites my current needs fine?  Though I use Win9x only for gaming, that's the same question everyone else should be asking themselves, even if they use it for productivity.  Now, for MS to continue being successful, what they have to do is somehow convince people that -- despite getting along fine as they are -- they *need* to upgrade.  In hard terms, what they need to do is convince people that the extra money they spend on WinXP -- and probably the extra money they spend on new hardware to make WinXP run decently -- is going to be made up for considering the TCO.  For the home user, I find that extremely hard to believe.

As hard as it is to believe that for the home user, it's even harder to believe it for the business user.  Now, some people will claim that WinXP has a better user interface, and I assume they're speaking about defaults.  Regarding the start-menu, I partially agree.  Placing it all the way in the screen-corner is better.  However, does it matter for any experienced user?  No.  People who use Windows regularly will use the Windows start-menu key, not the mouse to access the start-menu.  And of course, the way the start-menu has been de-logicalified in the default is not good.  So, the user interface issues are non-starters.  What about other areas?  Stability?  XP offers nothing over 2k.  Why should a corporation switch to something new and unproven, when what they have works fine?  Security?  XP is as weak as the rest of the products.  There was just a newspaper article discussing a bug in MS' IE and Outlook/Outlook Express which allows a web-site or e-mail to execute malicious code.  It's true that these aren't parts of the OS, but MS is selling an entire integrated solution with no choices, so they should be judged by that integrated solution.

Ok, now lets focus in on the Office products.  I believe that Office 98 also had the ability to copy & paste between tables and Excel.  Also, fyi, excel is a terrible place to print out tables from...for example, no super-scripts or sub-scripts in Excel are allowed.  For printing out tables, Excel just isn't as good or flexible as tables in Word.  Excel is just good for making graphs from tables (also, can't merge cells in Excel, but can merge them in word-tables, etc etc).  MS' HTML-converter is garbage and a half-completed idea at that.  Documents converted to HTML will still not display appropriately.  HTML was not meant to have such complicated document formatting, and any attempts to do so will result in poorly designed web-pages (any web-page where you can't resize the text is poorly designed).  If you need to share a word document with the world, it'd be better to put it on the web as a PDF-file.  Foreign character support is useless to the average user in the US, so, again, what does Word 2k offer over Word 98 that justifies all that money MS wants?  I would argue that maybe every four MS "upgrades" is actually worth the extra hundred bucks over one four issues back; but each sequential upgrade is a complete rip for the money they want.  Most of the improvements you spoke of in 2k over 98 were and 98 over 95 were just fixing things that were broken (table/spreadsheet integration between Word and Excel, for example).

Most of the upgrades MS offers over previous installments of their products are nothing more than insults to the consumer, becuase they are asking the consumer to spend an extra hundred bucks on something with new features that aren't usseful to the consumer, and worse this new product will perform worse than old ones -- take up more RAM, more HD-space, more CPU-time, etc.  MS and other corporations are going to run into these problems more and more -- program UI's and functionality eventually  reach plateaus.  Word processors are pretty much at the optimum right now, and so are UIs.  MS realizes this, which is why they are now trying to switch over to a rent-a-ware business model.  This is, of course, putting most businesses off.

This is the practical reason why I'm sticking with GNU/Linux for my productivity needs.  They don't feel the need to add new features to convince users to upgrade; they just deal with what's most important, and if there is a UI problem or functionality hole, they fix it.  But they don't try to invent new "user needs" to try to convince the user to upgrade.

The entire software business model for the proprietary world seems to be convincing people that they need to pay for upgrades when they don't, and that these upgrades will actually save them money in the long-run, which they won't.  Because these new more hardware-hungry programs are pushed on the end-user, it fits in nicely with hard-ware companies trying to convince home users that they really do *need* those 4GHz CPUs.

Thursday, March 20, 2003

Just me,

Why would FS / OSS shops maintain a distinction between new versions and "patches" to old ones?  Almost all of them are offering their software online for no cost, so there's no point.  The only reason why businesses have to maintain patches for old version of software is because otherwise customers would be pretty upset with them -- "What!  You're asking me to upgrade to your newest product and spend an extra $100 just to get a security patch!?"  In short, consumers would see it -- rightly -- as a scam for force upgrades.  I wouldn't be surprised if corporations start doing that, in fact.

FS / OSS projects usually do not charge money for new versions (though they can), so they integrate patches / fixes and upgrades.  Since upgrades -- at least for the apps I use -- don't add bloat and detract from performance, that's fine.  If there are some programs -- like KDE -- where upgrades seem to be always adding more bloat and making the product slower, in addition to providing new functionality and fixes, then people can choose differents apps.  How would maintaining and patching old versions help anything on the overall scale? It would simply mean more documentation work for the programmer (because he'd have to separately document various versions, or add "X for 2.1, Y for 2.3" in manuals, which would take away from programming to get real work done.

Thursday, March 20, 2003


In a business setting the maintainance and support costs often far outweigh the aquisition cost of a piece of software. Patches are not supposed to break applications, new versions often require adaptations. Even the most backward compatible focussed companies, such as MS, often have breaking changes in new releases.
Microsoft was rightly critisized in the past for exactly the problems it caused to administrators when they did not separate the new functionality out of the service releases. They are now better at this than they used to be.

I have seen problems in shops using certain OSS package, where upgrades to the latest and greates release was not an option because too much code would have to be rewritten, and yet they could not support the current product any longer for lack of essetial security updates. They decided to quit the product rather than starting their own infrastructure maintnance effort.

Let me try to give an example that hopefully flies below your ABM radar. If tomorrow the 2.6 kernel is released, how do you think the Linux community would feel if 2.4 support would be dropped at the same time? How many 2.2 kernels are still out there? Should they all be forced to accept 2.6?
We move the technology forward, and sometimes that means breaking with the past. A responsible company has a clear lifecycle roadmap for its products. Especialy for critical security issues, "that issue is solved in the new version" just does not cut it.

Just me (Sir to you)
Friday, March 21, 2003

*  Recent Topics

*  Fog Creek Home