Fog Creek Software
Discussion Board




Upgrade to the latest and greatest

From the H1B topic that bursted into flames Crusty Admin brought up an interesting point about software upgrades.

<quote>
I find it interesting that people that run typically UNIX apps - apache and the like go with "the if it's not broke don't fix it."  The converse seems to be true in the Windows world.  "If it's newer, it's better, upgrade now!"
</quote>

In fact Apache 1.x is the latest and the greatest. That probably sounds confusing. Let me explain myself. Disclaimer: I am grossly overgeneralizing.

In the UNIX world, especially the free one, there are generally two lines of products. The "Stable" old version (e.g. Apache 1.x, Debian Stable, FreeBSD 4.x) that does get updates, but only the security ones. Then there is the new "Testing" line (e.g. Apache 2.x, Debian Testing and Unstable, FreeBSD 5.x, Redhat Fedora) that get security updates together with new features. Conservative (read smart) system administrators will always run the boring outdated stuff. Outdated in features not security updates that is. No new features just means less headaches. And why would you install new features every month or so on a email, database or web server?

You probably guessed that I am still running Apache 1.x, MYSQL 3.x. and thoroughly hate Redhat for ending support for RH 7.x. Bye bye Red Hat. Hello Debian and FreeBSD.

In the Windows world software security upgrades generally come together with new features. The reason is simple, commercial companies make more money from customers that run the latest and greatest (read buy $$ upgrades), then from those that stick to the old stuff (read install 0$ service packs).

While from a technical and reliability point of view it makes perfect sense to keep those security updates coming for older software versions, it is the managers and stock holders that decide to stop that, thus forcing you to pay for the new upgrades.

Redhat as a commercial UNIX company followed the Windows rainbow, saw the pot of gold at the end and did the same. Lucky for them that purchase decisions are taken by marketing brochure reading managers and not the guy sweating it out in the server room.

Jan Derk
Thursday, January 29, 2004

Until recently, i was the product manager for a commerical software product. There basically are three or four groups that drive adding new features rather than just fixing bugs

The products that I work on are not standard IT products but engineering tools so it's not exactly analogous to the OPs comments. However it's close enough.

1) End users. The guys who drive your product every day. They don't make purchasing decisions but it's generally worth putting something in the release for them. Typically they want annoying bugs fixed and features to make their lives easier. Their requests are generally simple to fufull, but endless
2) Sales/Purchasing decision makers. These are two distinct groups, but the sales guys want the same things the people at the customer's companies want with money do. There may be some divergence as sales should be focusing on people who haven't bought in addition to keeping current customers happy. This group wants the big ticket items that will in theory improve productivity or connectivity to some enterprise software.
3) Management (if their smart). If you have smart managment, someones thinking of innovative features that will really make a difference in the market two, three, five years down the road.

In any case, it's not as simple as us wanting customers to buy upgrades. There is more than one motivation.

pdq
Thursday, January 29, 2004

Bursted?  Is that some kind of joke?

wtf-ed
Thursday, January 29, 2004

That is strange. My experiences are exactly the opposite. Granted, we use a lot of Microsoft stuff around here, and they might be better than average for the Windows world. Typically the Windows stuff is supprted for a very long time now. Heck, even Windows 98 has moved into the 7 year lifecycle now. Updates are free during the lifetime of the product.
Our Unix stuff OTOH comes with "support contracts" that typically cost more per year than the total Windows equivalent software. The support has become better over the last few years, but in the past it typically was "we don't tell you about updates and patches unless you tell us  what your problem is".
The few Linux setups around here seem hopeless wrt support. It's like they don't even have the concept of maintaining support. They are running an almost perpetual upgrade mill with constant breaking changes.
Maybe it is better if you run the "enterprise" versions instead of the free stuff, but then again, those are far more expensive in all respects than the Windows equivalents, so what is the point?

Just me (Sir to you)
Friday, January 30, 2004

My experience is that this has nothing to do with what platform you run, and is instead all about what you do.

If your job is to GET things running, you want to upgrade, so you have the latest and greatest stuff and can do more. This is mostly managers and developers, who tend to like Windows.

If your job is to KEEP things running, you want to keep what you have, so you don't break anything that works. This is mostly administrative and support people, who tend to like Linux.

There are plenty of people in admin and support who use NT 4, and plenty of people in development who use the latest and greatest Linux. I'm one of those get-running people, so I'm always running the latest Windows and Office and LAMP stack on my desktop. But when I work on our public server, I'm in a keep-running role, so we're a couple versions further behind on everything.

Caliban Tiresias Darklock
Saturday, January 31, 2004

*  Recent Topics

*  Fog Creek Home