|
Letters to the Editor -
Hi Peggy, I very much appreciated your write-up in EDA Confidential, and how you connected the standards panel you chaired at the Synopsys Interoperability Forum to Jason Hunter's talk at the Software Developers Forum. I don't know, though, that I would connect open standards with open source quite so directly. I think they're more orthogonal concepts. Open standards are about two things. At a technical level, they're about decomposing a problem in a way that hides implementation: "I don't care about how your black box works, and you don't care about how mine works, but if we can agree on the external behaviors, we can interoperate." Encapsulation is a cardinal engineering mechanism for managing complexity. At a political level, it's about "co-optetion." In Sun's phrasing, cooperate on the standards, compete on the implementation. The standards wars, then, are about seizing the high ground: "If my way of doing things becomes the standard, then my competitors are at a disadvantage." In short, standards are a way to agree about how to disagree. Open source, on the other hand, is about implementation. It's about building what's inside the black box. Yes, there is competition in the open source world, but it's more competition in the sense of academia: Let the better idea win. But open source does have a way of creating de facto standards that become de jure standards. As a friend of mine, Bill Grosso, put it: He's not particularly an open source guy, but he finds that he looks to the open source world as the place where the agenda is set. http://www.oreillynet.com/pub/wlg/5560 To analyze the relationship between open standards and open source, consider the following matrix (my apologies if it's mangled by the formatting of e-mail):
--------------------------- Open --------------- Proprietary -------- Standard .................... HTML ...................... NETBIOS Open Source ............... Apache ..................... SAMBA Closed Source ............... IIS ....................... W2K Server Apache = The dominant web server, managed by the Apache Software Foundation IIS = Microsoft Internet Information Server, i.e. their web server W2K Server = Microsoft's file server Samba = open source "reverse engineering" of Windows file server In the boxes, I've listed implementations that are emblematic of each of the standards/source pairs. What observations can we make from this matrix? 1) Microsoft tries to use it's market clout to "extend" the HTML/HTTP standard to lock users into its proprietary implementation. 2) Proprietary standards tend to be obscure. They force competitors to play on your ground. The implementers of Samba needed to "figure out" how the Server Message Block (SMB) protocol really worked. There is though, I think, a connection between open standards and open source. It has to do with the fact that constructing open source implementations tends to lead to open standards. It's more a social/sociological/political connection than it is a technical one. And therein, we come to the central point of Jason's talk: At its core, open source is about social structures, not technology. Lastly, forking is a central problem in the open source world. It's not done lightly, and I think that the time limitations placed on Jason's talk meant he had to gloss over the hurdles, both social and technical, to the process of forking. For an excellent discussion of forking and open source in general, see Eric Raymond's book, "The Cathedral and the Bazaar." Regards,Sanford Rockowitz
Peggy, I've found a number of copy edit errors in your articles. Perhaps you should try to take more time before posting your text. Signed, Anon
Dear Anon, Luckily you're not my boss, because if you were, I'd have to resign and give you a poke in the eye and a punch in the snoot on my way out the door. Instead, I'll just cancel your subscription to EDA Confidential. I'll be sure to see that you get a full refund. Warm regards, Peggy Aycinena *************
|