PDA

View Full Version : Article: When Good Design => Bad Product


Steve Mehs
03-20-04, 12:09 PM
---
This article was posted at AVS today, it’s from December, but still it makes a good read. I know it’s a long article, but if you skip around make sure to read the Final Thoughts at the end.
---

What a strange situation. You take a mediocre product and rework the design to make it better. Your design is a success, by any reasonable measure, but the resulting new release is actually worse. You redouble your efforts and matters become untenable. It doesn’t matter how brilliant and effective your designs, the more they improve the product, the less usable the product becomes.

What could cause such a situation? Industrial sabotage? A rip in the seam of the universe? No, poor quality assurance (QA) procedures.

Dish Network, with its digital recorder/receivers, has shown evidence of this phenomenon for several years now, with no end in sight. They have recently released version 115 of their software for their model 721, and it appears to have almost as many bugs as their very first release. These bugs are not subtle. They are easily reproduced and, in a few cases, fatal. Their genesis, in several instances, appears to be a reworking of the interaction design of existing features. The new designs are a great improvement over what came before—or would be if they worked.

MORE (http://www.asktog.com/columns/059QualityAssurance.html)

amit5roy5
03-23-04, 11:59 PM
interesting

Mark Lamutt
03-24-04, 07:13 AM
Sums it up pretty well I think...

FritzM
03-24-04, 10:19 AM
Brings to mind an old corollary to Murphy's Law -- Once something's screwed up, it stays screwed up.

tm22721
03-30-04, 05:59 AM
Nope it's not a tear in the universe.

Traditional software design methodolgy breaks down as complexity increases beyond a threshold that varies with an organization's size, experience and resources. This level has ben breached (in the case of E*) by almost all of their PVR efforts.

Bob Haller
03-30-04, 07:04 AM
His comments mirror mine. Thats why I have become so negative on E. Its sloppiness at its worst, or carelesness. Sad to say the same problems appear to effect the 921 rollout.

These things are preventable if charlie & company did their job better.

rdopso
03-30-04, 09:31 AM
Can someone send the article to Charlie?

Bob Haller
03-30-04, 09:39 AM
I want to thank the original poster. I am featuring this article on my newsgroup posting, since it shows I am NOT a troll and my probems arent just me.

Charlie is aware of the situation but doesnt care:( For HIM it the most efficent way to do business:(

if posters here quit covering for stuff like this charlies internet based sales would plummit and he would be forced to fix things.

Steve Mehs
03-30-04, 08:07 PM
Credit goes to AJF from AVS for diging this up.

TomCat
03-31-04, 08:39 PM
It sounds like a tear only in the writer's personal universe.

Where are the "facts" supporting the allegation that there are still a lot of bugs in the OS? Without those facts, this is just another anecdotal report of one user's experience.

My anecdotal evidence is exactly the opposite...slow but steady improvement since the 501 was released in April 2000...and that is just as valid so both reports are effectively cancelled out. While not Replay-rock steady, stability became a non-factor well over a year ago for me, and if the forums are any indication, for most owners. There's where the real evidence is.

Don't believe everything you read. A bigger soapbox does not imply accuracy even in the least. This writer is full of what the birds eat and only does a disservice to all of us.

Bob Haller
03-31-04, 09:41 PM
the software issues are with the 721 and they still exist.

willy
04-01-04, 08:51 AM
I gotta agree with the original poster's blurb also.

This is what happens when a company gets too big for its britches and bites off way more than it can chew.

If you start with a bad, or even mediocre design and spec, that itself is not fatal. What can be fatal IMO is the way you handle it. If you take the time to think out the solutions and possible ramifications, and do extensive regression testing, things can work out.

E*'s solution is to quickly put on a band-aid to fix the issue. If another issue pops up as a result, throw on another bandaid. And so on and so on.