Avadis “Avie” Tevanian, Apple Computer’s Chief Technology Officer, is leaving the firm to “pursue other interests,” the company confirmed Monday. Tevanian, who came to Apple in 1997 from his previous post working for Steve Jobs at NeXT, played a key role in developing Mac OS X, the company’s widely adored operating system. Although Apple wouldn’t say what Tevanian’s plans are, Vice President of Worldwide Corporate Communications Katie Cotton wrote in an e-mail that, “He plans to take some time off in the interim. He hasn’t left yet, though. His last official day is March 31.”
Avie Tevanian Leaving Apple
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Maybe *NOW* someone’s allowed to take care of the Finder?
Appleinsider posted some neat diagrams and info on Leopard’s finder that gave to the patent people.
http://www.appleinsider.com/article.php?id=1593
it is pretty cool.
Avie was a kernel dev, mainly. He can’t be blamed for Mac OS X’s ‘infamous’ Finder.
I guess you could say, though, that “his [kernel] work here is done”. Tiger is fast, pretty, and woo. Now, with Spotlight, Apple can focus on giving us a new, first-class-object destop, and the Finder, old-style, can die the death it richly deserves.
Tevanian used to be senior vice president of software engineering until 2003, and I’m sure he had something to do with the new Finder. Here’s a great article about him:
http://daringfireball.net/2003/07/the_good_the_bad_and_the_avie
I just read your link– I must have missed it in 2003. It is an EXCELLENT read. Maybe things are looking up, afterall.
No but he can be blamed for the kernel! Of all the big operating systems out there Mac OS X has the WORST internals by far. Slow, bloated and oh so arcane.
The most hated thing in Mac OS X. I don’t mind it _that_ much. The only real thing I hope they do to it is make it multithreaded. I don’t want to have to force kill it every time there is some kind of long network time out or something. I want it to always be responsive.
Oh and I want the option to stick a refresh button on the toolbar (for network shares).
Tevanian is probably the best OS guy on the planet. This will be a great loss, but I guess the “heavy lifting” is done for OS X. Maybe new voices can take it a new direction.
I think the best OS guy is Bill Joy.
Just my two cents 🙂
Oh, I didn’t know about TN2034.
Anybody that dogmatic really should have to go. Still, Apple made partial penance for its metadata crimes in Tiger (Extended Attributes/Spotlight/Uniform Type Identifiers) but they have some way to go.
Maybe going to help Haiku or Linux?
CTO’s do not leave a job like that to help OSS projects unless they are retiring and want to have something to keep their time (or he is trying to create a start-up using Linux… did not work out to well for easel)
Tevanian is Chief SOFTWARE Technology Officer, not CTO. It’s a more specific and less operative role.
Avie wasn’t anymore the responsable of Mac OS X development since 2003, it’s Bertrand Serlet and under his guide came out Panther and Tiger.
MacUsers are in good hands! 🙂
P.S. And if Apple will change the kernel? Anyone said Solaris? 😀
Edited 2006-03-28 16:32
To bad he is leaving Apple. Not to many other places to go from there…. unless he heads over to google… hmmm
It’s clear none of you worked at NeXT or Apple because the character assasinations are just that, assinine. Gropw the hell up and recognize the FINDER is an Apple compromise, not a NeXT design. Finder was hated by everyone inside of NeXT engineering.
To bad he’s leaving!
Just out of curiosity, I wanted to ask the question… what do so many find so wrong with the finder?
I’m a Mac user from way-back when (I loved playing Lunatic Fringe in After Dark on System 7), and been up-to-date with OS X since 10.1, and I really just don’t have a problem with it. Personally, I think it keeps getting better and better, and now going back to 9.22 or *shudder* Windows seems a pain in the rear. Aside from a missing tree-view (which I personally don’t like), I don’t see any real missing functionality.
What are some of the things you don’t like about it?
(this is an honest question, not an invitation to flame-bait, by the way)
Beats me. I think some people believe it should have better threading.
One thing I really miss is the spring loaded folder behaviour. The way it works in OS X is a mere shadow of its former self.
Avadis Tevanian is Armenian name; which is same heritage as mine. Makes me very proud of my blood.
heck yeah…. me too 🙂
From what I read in the past this guy is one of the reasons why os x was so slow before. This guy wouldn’t even consider using l4 for os x because he did not want to give up on mach.
No, Tevanian wouldn’t consider choosing Apple’s PPC-specific kernel (nuKernel) over the one he co-wrote in graduate school (on other architecture).
In 2003 it was all well and good for Daring Fireball to mock this decision; in hindsight it seems to have worked very well for Apple.
No, Tevanian wouldn’t consider choosing Apple’s PPC-specific kernel (nuKernel) over the one he co-wrote in graduate school (on other architecture).
Not just that. It was a choice between Mach, which was already in NeXTStep, and which he and his team at NeXT already had a lot of experience with, and nuKernel. Remember, when Apple bought NeXT, most of the top people were replaced with NeXT people. It makes a lot of sense for them to stick with what they knew.
In 2003 it was all well and good for Daring Fireball to mock this decision; in hindsight it seems to have worked very well for Apple.
Apple really didn’t have a choice. They already had an enormous task ahead of them, and decoupling NeXTStep from Mach would’ve taken an enormous amount of time. They already cut corners like sticking to Mach-O instead of moving to ELF, to save six months of development time, so a project that probably would’ve taken a year or two was definitely out of the question.
Ultimately, Apple knew that OS performance really wasn’t a big deal to their target market. On a workstation, as long as you’ve got fast streaming writes to disk and low latency, other things are pretty much not a big deal. Unlike server apps, workstation applications generally don’t spend a lot of time in the kernel, and the bottlenecks tend to be in userspace code. Even today, if faced with the choice of spending a year ripping out Mach and replacing it with something else, Apple would probably be better served by improving the OpenGL stack or something instead.