Hi Claus,
first, it's really appreciated that we have such a communication from your side.
At Wed, 22 Aug 2007 17:22:05 +0200, Claus Riethmueller wrote:
Hi James,
please let me say first that I am not intending this as any personal comment as I am pretty sure you are not the only person with similar thoughts about the discussion.
Yet, I have to say that it is exactly the somewhat ignorant attitude that turns us and quite a number of other vendors "off" in supporting the ALSA developer community better. It is what I have refered to in my email to Piotr that he cut & pasted to this mailing list (without asking for permission btw.). I have been following the developments at ALSA since many years now and worked with a number of different developers for different products to get them working properly. Many of our products are supported just fine under ALSA and both the developers and the users seemed to appreciate that fact. Unfortunatly your comments show your obvious lack of understanding about how some companies are working in this industry and what sort of approach is needed by you guys to get better support from us. This is why I used the term "culture crash" in my email.
I know this kind of conflicts well from my long experience, and I'm sure that James also knows. The fact is, that there are minimum requirements for the driver development, and it couldn't be seen from the cited mail.
So, basically, ESI cannot help us. I am amazed that they can write any sort of driver without some sort of datasheet.
Well ... frankly, it should not be your concern on how we are developing drivers for our hardware designs (which are ours). What you wrote is a cheap comment and I do not see how it could be an appropriate comment if you look at everything what I wrote and not just the introduction. The way (format, style and most important in our case: language) we store our internal company data is no ones concern except mine and the concern of my employees.
As James mentioned, the datasheet is vial. "SOME SORT OF" datasheet is. Yes, this can be any document. It can be even a document under NDA, at least, for the beginning of the development. (But they are often released later in the public place after rewrite.)
But, when the information is given under NDA, the most important thing is that we are allowed to write the driver without *any* restriction. It means that even all information in the datasheet could be documented as comments. (Of course, we focus only on the technical stuff, so there haven't been problems about it.) In other words, for the open-source development, NDA is nothing but a formal paper work :) After all, it'll be a question of trust.
Oh, BTW, please don't misinterpret: NDA is the last choice. The public document from the beginning is the very best.
(snip)
Our engineers are all happy to answer questions and provide assistance and help if required. As I said, we are happy to be in contact with an individual developer who is actually doing the development. It is a lot less time consuming and for us a lot more simple (actually, I believe it is more simple for the developer on the ALSA end as well but I realize some people might disagree here).
Some people are already interested in the development, so it would work actually if they can receive an actual hardware for testing (yes, the test hardware is another vital factor).
The development (driver writing) itself doesn't have to be driven by the core ALSA developers. They, including me, are often too busy by other tasks. So, other interested developers could work better on it, I guess. But, it'll be anyway nice to keep in touch with ALSA developers (i.e. adding them to Cc in mails). This will make the later maintenance job a lot easier.
thanks,
Takashi