Why does PC computing still suck?

<disclaimer>This post may seem to target Microsoft unnecessarily, but that is mainly because this is where the majority of PC use occurs.</disclaimer>

Please read the following 3 articles, written by a newspaper columnist about his experience using PCs on the corporate network. (These are from his personal blog). Some of you who have seen me link to James Lileks website before. I personally think his writing is hilarious, but funny or not, these stories are quite telling:

http://www.lileks.com/bleats/archiv...202.html#101002 (start at the 5th paragraph)

http://www.lileks.com/bleats/archiv...203.html#102302 (start at the third paragraph)

http://www.lileks.com/bleats/archiv...201.html#110702 (start at the 5th paragraph)

THIS IS THE TYPICAL PC USER EXPERIENCE.

I consider these as prime examples of how bad PC (personal computing) still is, more than 20 years after the invention of the personal computer. In fact, if anything, it seems we've gone in reverse, in some ways. Software is so bloated these days that the typical user doesn't even use 1/100th of the features. The "power user" might use 1/20th of the features. (Word is a perfect example).

Mr. Lileks' experience is typical of the resignation many people feel at the unpredictable and nonsensical things they have to deal with. "Resignation" is the correct word here. It's not that they think its the best choice; they have no choice. I firmly believe that the real reason Windows became so popular was not because of its user interface, or that "grandma" can install it (she can't). The real reason is because Windows made it easy for just about anyone to become a developer, or at least to think that they have become a developer. Microsoft had it right. Make a series of tools that pretend to make programming child's play. Who cares if these tools introduce all kinds of interdependency and versioning problems? Who cares if these tools produce programs with nowhere near the robustness of C, C++, TCL/Tk, etc... No, these problems don't hurt the developers at all. The developers get a reduced product delivery time, and a radically inflated product support contract, just to keep the program running, and surviving through DLL hell.

Yes, Windows can run over 100,000 applications. So what. This is simply too much. There is so much junky software out in Windows-land that it is hard to find the decent stuff. I don't say this as a grouchy old Unix guy. I have spent my time supporting Windows systems in small and medium-sized businesses, and my experience is consistently that Unix (FreeBSD or Linux) would be easier to support in almost every case, as long as the company wasn't limited to a Windows-only version of the software they depended on. Industry-specific software is the worst. I remember one insurance compay customer who had to run through the main CD install, and a whole tray of floppies containing patches and upgrades, every time she needed to install a new workstation, and if certain things weren't done in just the right order, kablooey!! With each new release, invariably, something wouldn't work right, and we would have to wait on the phone for over an hour for someone to finally tell us to just reinstall it, and it should work. FInally we would discover that the documentation was wrong, or that some DLL was newer than expected, etc...

I'm not trying to say I would put Linux into the hands of the average Windows owner and say "go to town". (Actually, in some cases, it is almost this easy). My point is, the reason Windows is the de facto standard has almost nothing to do with useability, and more to do with Microsoft's timing on the market, and the plethora of 'fast-food' programming that they encouraged.

Yes, a lot of this is indicative of how much computing in general sucks, not just Windows computing, but it seems to me these end-user-land problems are always worse with Windows.

So, rant over. What are we gonna do about it?

Thin clients are the way to go.

Client-server is an overcomplex, antiquated relic.

I have spent quite a few late nights overwriting the wrong registry key, or the wrong DLL. With UNIX systems (I don't include Linux in this group -- I'm talking HP, Sun, AIX, BSD, SCO), basically everything is confined to the scripts and executables in one nice directory. Very clean, very tight, makes upgrading and reinstallation a breeze.


I still think XP is the best desktop OS out there, but the thick client model itself needs to be destroyed altogether.

If you can marry the high availability/redundancy that many pc workstations gives, with the central administration of a mainframe, that's the way to go.

All class Rycamor and I dig the articles, rofl There is no doubt something needs to be done, M$ have almost forced us into their crap, over priced products.

First, a rant about the evils of M$ :roll-eyes: They have gained their market share through copying their competitors, buying them out and even using illegal market practices (using a market share in one market to gain a share in the other, using Windows to sell IE). Not only do they suck, but they don't care about the consumer, they are unethical, and they certainly don't make much contribution at the charity level. Thats a lot of thumbs down

So how can we break M$'s stranglehold on James User's PC mindspace? To answer this question we need to know two things: why people stick with M$; and how they can migrate.

Why stick with Windoze?

*Windows only software
What stops us going to all linux is those applications that don't run on linux. I use Flash MX from time to time and there is no linux version. I unfortunately think that macromedia and microsoft are bathroom buddies, and I saw macormedia totally crushed openlgx, which would have even an open source version of flash, just like M$ would crush linux if it could. So it is tricky

*Familiarity
People tend to stick with what's familiar (just look at Friends). To me, a big part of why windows is easier to use is it's GUI base. That would suggest that lindows and Redhat are probably the open source OS's with the best chance of competing.

*Compatibility with other users
Being able to read documents from other people or other organisations it really important. This can really only be solved by the adoption of standards, rather than proprietary document formats.


How to make it easy to migrate to open source OS
Part of the problem is that Windows, Office and IE all come bundled together. Switching all at once is a bit like being thrown into a mountain stream (invigorating if you can handle it, but too much for most people). The migration path needs to be as painless as possible for people.

People don't choose a software platform for it's own sake. They choose it because of what it lets them do (what software is available, ease of use, performance, security etc). If users move away from office and IE, then this weakens the Windows platform as a whole.

IMHO our best chance for a migration strategy is:
*Change to Open Office
*Change to Mozilla
*Change to Linux

I'm not sure if OpenOffice has a good e-mail client (I still use Express ), but this is probably a 'must have' as well.

*******************************************
An Ode to the virtues of OpenOffice and XML
*******************************************

<Ode>
I'm successfully using OpenOffice. For me the only problem is that OpenOffice can't extract the images from word properly when converting documents. Aside from that it works really well for me. If you save as writer format (web page) you get nicely formatted XML, properly referencing your images, which are copied nicely into your folder. This is great! The benefits are: cheaper, works better, and saves data in native XML format.

Storing data in XML format is a big deal. You can't do anything with a word document, it is massively oversized heap of steaming dung. Everything is embedded, However, XML is text, (not as efficient as binary, but light years ahead of word), it is flexible, easy to translate, accessible (e.g. easy to read by screen readers). XML is flexible and portable. For one, it's easy to move to the web.

It can form the basis for a document management system as well. Combined with an XML schema we have a method of defining document types, with XML fields/nodes/attributes. Lets say I want save my clients, I've got <name>. <email>, <phone>, <address> and maybe <desc>. Lets say the user wants to format the description, I simply allow html format tags in here. Great! Then I have either a php template (for web) or xslt stylesheet (whatever other format) to format this data, combined with an XML editor, combined with an XML document management system, and I've provided the core functionality for organisations to manage, author, publish their data in any format they want.
</Ode>

So what does this have to do with M$? Well there is an XML editor for word, WorX, but it costs and so does Word. If people want XML, then there is a chance we can convince them to go with OpenOffice. Now remember OpenOffice covers all the main things users do on their computer, this is the first big step in getting away from M$.

On to Mozilla. I can give you another 101 good reasons why you should switch.
http://www.xulplanet.com/ndeakin/arts/reasons.html

In addition, not only do we have OpenOffice providing us with XML, but we also have a very cool XML browser based editor, bitflux (www.bitfluxeditor.org). Although IE also allows browser based editing, you're not really doing inside the browser. It requires the client to be Windows.

If we can get users away from office and IE, then we've got a pretty good chance of getting them to use *NIX, providing it is GUI based and usable. I don't know how close it is, but I think redhat it pretty good in terms of its GUI.

So thats the plan, but the obstacles are everywhere. No one wants to change. When I left the local university to set up a business, they were considering forcing all users to go into windows to make support easier! Hell if the universities are acting like this, most businesses are probably like that. OpenOffice and mozilla/netscape as developers we have a voice that managers must listen to.

Then there's linux. If a painless install (as painless as windows, anyway) can be produced, then all of the advantages of speed, cost, support, security etc I'm sure will shine through

Whew Its the weekend now, so I'm outta here.

Awesome Z.

I agree partially with Ted Striker. Thin clients are a great solution to the business network. Centralized administration, etc... More and more corporations are taking the plunge to install Linux as a thin-client network, and generally the case histories are very fun reading.

But thin clients are not an answer to the typical home user. I am personally still bemoaning the death of BEOS, which I thought had potential to be the best personal-user OS ever made. In fact, even now it still is better than anything Microsoft or Apple has put out since.

Quoting from Mr. Lileks' writings: "The botherations we put up just to use these machines boggles the mind."

Amen, brother...

I spent several years helping people troubleshoot and set up Windows. The number one experience I take away from it is: it is simply incredible the amount of trouble people go through just to continue using their Windows PCs. I honestly think people spend almost as much time fussing with their PCs as they do actually using them productively. It is an entrenched problem in our society now. I don't know what the solution is, but it seems to me that neither does Microsoft, in spite of all their expensive research projects.

It's frustrating: Microsoft made a series of very clever business moves, but almost every one was a bad technical move. They started with DOS, which was a simplistic hack of C/PM ("Quick and Dirty Operating System" was its original name.). But at least DOS was fairly simple, and behaved as expected for the most part. Then, they began building a series of front ends. It's kind of like trying to build a mansion by starting with an old wooden shack for the foundation, then adding a few rooms, then a facade, then trying to remove the shack foundation and replace it with a slightly better shack, without disturbing the rest of the house.

Awesome Z ,

I agree with just about everything except the "painless install" bit. Every time I install Windows (and the applications needed to actually do something with it), I know I am in for a wasted day of my life. Yes, with luck, you can just cruise through the standard default install, but then what? You still have to apply several OS patches, then apply the IE security patches, then install ZoneAlarm, then an antivirus program, chasing down device drivers, (rebooting and finger-tapping at each flip of a bit) and then you can finally get down to installing your apps. At this point, if you are lucky, your apps just work, but often there are other problems to resolve.

Doing a standard default install of RedHat, Mandrake, Suse, etc... (not that I use these) can take about the same time as a default Windows install, but at the end of that you have a fully workable system, complete with apps, decent security, games, productivity software, etc... For a geek like you and me, FreeBSD provides just as fast an install over the web, complete with up-to-date software and patches anytime you want.

The only difficult areas are the drivers and configuration for "extended" devices, such as video capture cards, TV receivers, and other specialized hardware. This is only because the makers don't provide Linux/Unix drivers for their equipment. Usually a driver can be found somewhere, but it takes reading, and tweaking to get it right.

One interesting Linux distribution these days is Xandros. Apparently they have beaten Lindows at the game of Windows app support. Let's see how that turns out. I think the company that truly solves that problem is going to make a boatload.

rycamor,

Has BeOS officially died? I thought Palm bought it right? Did they kill it?

Yeah I'd have to agree that my thin clients work best for business, it might make sense for alot of use to have thick clients at home, for more freedom, flexibility, and privacy.

Citrix was all the rage in 1999, but the hosted apps technology seems to have stalled, probably because it takes a killer server hardware resources to host all of those simultaneous apps that it just wouldn't be cost effective yet. Then there is also the bandwidth issue.

Now the focus in on web services.

Though if a company could make a profit out of offering thin client applications to the home user, that would rock. Yahoo! is starting to do this with some of their games on demand, it's interesting to see them continuously moving forward:

http://gamesondemand.yahoo.com/play

EDIT: Hey, i just downloaded their games on demand app, it's about 3.5 MB. For each game you want to play, you download about 50 MB of a cache, and then the rest comes over the network (was seamless for me). Then it plays just like a normal game. The 4x4 racing game was pretty fun but I felt my Pitt Manuever should have worked more effectively.

Pretty impressive so far!

Up early, can't sleep, so I thought I post another episode of "Why does PC computing still suck?". In this blog post, Lileks muses on a particularly irritating piece of consumer-level software. I wish every useability group in the world would study his writing. Search for "PC Help" in this page.

Really, the stupidity of some software is just amazing. As a punishment, each of those developers (and especially the project manager) should be beat over the head with the 3-volume set of TAOCP, then they should be forced to read the entire set, first in reverse, then forward, doing all the exercises on paper each way.

Me.rant.setMode("off);

sigh...

Aside from this guys bias for MAC's.

[my rant]

>>Jean-Charles said that the program made him feel stupid, because he couldn’t figure out the simplest tasks, and I reassured him that this was hardly the case. If the software makes you feel Geefish, it’s the software’s fault. Video-editing on this level shouldn’t be hard. It’s checkers. If the program makes you feel as though you’re playing chess in the dark with oven mitts, blame the idiots who wrote the program.<<

This is part of the problem. Bad software. Why?

Can't any developers write good code?

Or is it that the UI is not considered an important piece of the app?

I write UI's for new engineering apps. I have a non PC user test them. I tell her what the end product is and she tells me if the program is easy and intuitive to use.

Then my boss and the customers say NO. Change it, make it more complicated because they use PC's all the time. I view it like Heisenburg's Uncertanty Principle, (the closer you are tot the problem the harder it is to be objective)

Change this and change that, stick in 50 more features to do tasks that no one else wants, but don't give me or the tester time to de-bug them.


>>I describe the problem, and they say they’ve been hearing complaints about this. They have no idea why it’s happening. It just is. I watch as the tech remotely trashes the program and its brood of ini files, then reinstalls it.<<

So why did the help desk not find and fix this problem before they upgraded everybodies PC?


Did they have the resources/time to do these tests?
Did their management leave these tests to the last minute or not consider them important enough to do?

My boss likes to debug my app minutes before an update is issued to the client as he considers the UI client the least important (compared with the hardware, server and image processing).

We have a major update today at noon, he started testing my app at 4pm yesterday.

[/my rant]

Ahhhhhhhhhh, that feels better

I get to see a lot of computer users in my biz (computer store owner, and web developer). I think a lot of the blame for software problems definitely needs to rest on the shoulders of developers for writing poor software, but the number one culprit for problems are the users.

I see the way people go click crazy on their computers and wreck things, they turn off antivirus software but open every attachment they get, they try to run the latest software on a computer built in the stoneages, they don't bother to follow/read the simplest of instructions or use any amount of common sense, I could go on for hours.

When things start crashing they complain that their computer is junk, or windows/software doesn't work right, or something is too hard, when they are the problem.

Blame should not be placed squarely on M$, Linux, or developers, it also rests on users.