What I see in this guy's demeanor is simply a short fuse and unjustifiable angst for not getting his way.
Some of what he says is correct- dealing with dev. rel. can be frustrating, especially when you are on an already delayed schedule and can't get a solid answer or resolution to something that is going to put you further in a hole.
I haven't dealt with dev. rel. from the leaders recently, but had similar experiences back in the V1->V3, TNT->Geforce256 days. It was why I often sided with 3dfx from a dev. rel. standpoint, although this went to hell in a handbasket shortly after the V5.
What a developer really *wants* is concise, realistic information concerning issues and not fluff. If I found a problem in a driver, within 24 hours I got a definate answer from 3dfx. Even if it wasnt what I wanted to hear, they sent me the direct answer that made my decision making a no brainer. It wasn't uncommon to get "Yes, it's broken, but the driver team has no intentions or schedule of fixing in the forseeable future" which equated to "start the 3 week work-around schedule" rather than "hop on other content and wait for patch that may never come." This surely beats the limbo of there being a chance of it being fixed in 3-4 weeks, or worse yet a committment to the fix in 3-4 weeks then still not having it fixed 6 weeks later.
With some IHV's, the "run-around" technique is golden. NVIDIA did this constantly with never a confirmation of an issue, but instead an "alternate preferred method to try" which were, in turn, as equally broken. I've heard from collegues they no longer participate in this wild goose chase method of developer support, but for the longest time it was very frustrating to have to drastically (and unelegantly) recode portions just to buy time around a driver bug that was still revealed as broken regardless of approach or method. It left the developer more weeks behind with no resolution, and the firm understanding the the dev. rel. representative knew this from the start but just wanted to buy time for their inept driver developers.
Our in-house CS team had management training that preached this same approach and I vehemently fought against it. The whole old-school method of "giving the client something to do unrelated to the problem to buy time" is the worst way to run customer relations.
While this Derek fellow makes some rather pointed accusations, there is little to no support to define these as real or self inflicted so it's impossible to ascertain the truth behind his claims. For the most part, I see his complaints being ones of frustration and they really don't match ATI's efforts in the past few months. I am definately seeing even *public* confirmations of bugs, issues and even (somewhat loose) committments to fixes at the consumer level. This is pretty unique to 3d hardware providers as when was the last time you saw another IHV come forward and state "XYZ game is confirmed to have lock-up problems" or "wbuffer is totally broken at this stage" etc.etc. ? Consumers love this stuff as it saves them time whacking their registry, re-installing or following the advice of morons that don't even have the hardware claiming the issue is non-issue. It's dangerous because the competition can abuse this (as we are already seeing), but puts the consumer's needs first.
Most IHV's only lay claim to *fixes* after they have been resolved and avoid stating flaws prior to fix. (i.e. VZ.ZZ - fixed issue with Z clipping in game XYZ, etc.etc.).
I like the new angle ATI is taking and I believe it will work well with consumers. A little initial frustration knowing there are existing bugs, but once they realize how convenient it is to *know* what exactly is broken, they will appreciate the time saved from this method.
Just my $0.02,
-Shark
