Xilinx documentation typos

Since you continue to refer to my original post as "nasty" and now "agressive", I ask you once again to explain exactly what part of that post you found offensive.

I am hardly the first person on the newsgroup to express exactly these same opinions about the usefulness of WebCases, and your documentation system's handling of both customer feedback and long-known, yet poorly documented, problems.

My comments are based on 20 years of experience using Xilinx parts; during that time I have attempted to have Xilinx correct serious flaws in their documentation and/or software on several occasions, with no success, as described in my original post.

Brian

Reply to
Brian Davis
Loading thread data ...

You might not want to go down this road. The Xilinx reps in this group are a bit quick to assume that others are posting with an attitude all the while giving all the attitude in the world in their own posts. Peter is actually one of the nicer guys, but he has his moments as you have found. I am of the opinion that it may be a bit of a cultural thing at the company. They seem to have developed a corporate opinion over the years that they are the best and are better off being defensive about their faults. Of course this is not accurate as a blanket statement. There are many times when they are very helpful. This is a comment about the "tone" of the replies you may get from time to time.

My point is that it is not likely to be worth the effort to try calling them out on the "tone" (or even substance) of their posts.

Reply to
rickman

this group

Reply to
Peter Alfke

Thanks, Peter. It seems the apparant name calling was the problem. Brian - you can fill us in on what you actually meant if it's different from Peter's interpretation. Personally, I saw other interpretations and saw the text as trying "to be cute, but [he] blew it."

The original post is pertinent. I stopped trying to submit webcases a while back, instead calling straight to the Xilinx hotline to kickstart the issue rather than dealing with the typical email response of "we need more information" even when extreme care was taken to provide all needed information. I was even called by someone from Xilinx asking about my use of the hotline versus webcase; I fed back the information I felt pertinent.

I like the idea of a "mystery shopper" approach where Xilinx can do an informal audit of their own procedures by submitting cases of their own. It may sound at first listen like it would be a waste of resources, having a CAE doing work that results in no actual help to the customer. But it could be a serious help to the customer if it exposes the limited effectivity (at times) of the WebCase process. It appears that we can't submit bugs for CRs unless we submit our design since software likes to have cases that can reproduce the bug to prove a bug fix. My company is hesitant about shipping out source code in spite of the bidirectional NDAs in place.

Webcase isn't (or wasn't when I used it) a terribly effective way of communicating an issue no matter how well considered in the original submission. Even if 60% of webcases succeed in providing a timely accurate response, the experience from the other 40% swamps the engineers memory.

- John_H

Reply to
John_H

John,

Peter and I are very serious about improving the customer experience.

After all, that is really all that counts: happy customers all designing commercially successful products with Xilinx FPGAs.

This dialog is providing for some real creative thinking, so please let it continue.

Just one comment, for every case that is resolved in a un-satisfactory way, it takes more than ten cases with the same engineer to repair the 'damage'. The good news is that we are doing much much better than 9 in

10 cases successfully concluded, but all it takes is a few to cause a lot of anguish (and unhappy postings to a newsgroup).

We don't even want a few cases to go unresolved, but realize realistically that some cases will be. Just have to make those as few as we possibly can, and be sure that we learn all the lessons we can from them.

Aust> Thanks, Peter. It seems the apparant name calling was the problem. Brian -

Reply to
Austin Lesea

What about the _really_ radical idea, of getting the "mystery shopper" to actually try and use the tools, and submit a real web case, based on whatever issues they encountered ? :) [that way, they do actually help the user base !]

-jg

Reply to
Jim Granville

Jim,

It may surprise you, but the FPGA Lab (which I manage) leads in the verification and characterization of new silicon.

In order to do this, we use the same software tools the customers do (the surprising part, perhaps?).

So, we typically file hundreds of change requests, and bugs per month when we first get the silicon back.

Yes, we are the first users, and no, we don't do just "goofy" special stuff, but are required to help get things like the PCIe designs working for the plug-fests and testing off-site, the DSP filters to work, the FIFO/BRAM to pack properly, measure the SRL16 power, and so on.

Next time you wonder "Hasn't anyone ever even tried this before!" the folks in my group probably have, and may have a bug fix request in for it.

If not, we probably just got lucky, and missed that particular case. We try to keep lots of old cases around for regression testing (just common good practice).

We are not alone, there is the systems engineering group, the applications group, product test group, the DSP group, the embedded systems folks, early access customers, design services group, all the FAE's that can't wait to see if the new part is a better fit for their customers, and a whole slew of other quality control systems in place.

Admittedly, all it takes is just one bug that isn't caught to ruin your day.

Aust> John_H wrote:

Reply to
Austin Lesea

That comment was not directed at, nor >

From my perspective, that is an accurate, albeit tongue-in-cheek, description of Austin's posting tactics in the past when I've posted detailed DCI/LVDS problem summaries to the newsgroup.

But there's no accounting for my sense of humor.

I've given up on them completely except for bugs I can't work around, or simple questions for which I can't locate the answer.

Each of us has bumped into differing subsets of all the sharp and pointy bits in the tools and documentation; if I choose to list some of those problems here, that does not mean I'm anti-Xilinx.

In comparison to WebCases and Xilinx website searches, once I've posted a problem summary, list of relevant Answer Records, or code example here to the newsgroup, I can easily find it again from any computer.

Brian

Reply to
Brian Davis

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.