Comments on: Apple’s new Mactel’s and UserAgents

By: James

James — Tue, 21 Jun 2005 03:00:40 +0000

Are Apple actually going to be using i686? Given Intel’s and Apple’s roadmaps, I had assumed that they’d be switching to 64-bit Intel chips rather than working in 32-bit mode.

That would give an architecture string of “x86-64″, “amd64″, “em64t”, “x64″ or whatever else it is being called these days.

By: Robert

Robert — Fri, 17 Jun 2005 19:14:10 +0000

@dave: the problem with Universal Binaries is the download size. It’s *much* larger. By deciding what the user really needs (rather than give both) there is a serious savings.

Why should the user download a much larger slower download when they only need half of it?

@Laurens Holst: I don’t really like Windows, but don’t ‘hate’ it to much either. My problem with the UA is that it’s the only thing we webmasters have to decide how to deliver content to the user. The user expects things to “just work”. It’s up to us to deliver the right content. If the UA is reliable and consistant, we can do this very well and provide a better experience to the end user.

Nobody likes to decide which download they need from a giant list of cyptic acronyms and technical jargon. That’s why being able to detect what the user needs is important. So we can deliver the best experience with no effort on their part.

By: Laurens Holst

Laurens Holst — Fri, 17 Jun 2005 17:15:02 +0000

Do like Windows? Didn’t the IEBlog spent time to shed light on IE7’s UA string recently? And got highly critisised by many for that, as focusing too much on UA strings!

~Grauw

By: dave

dave — Fri, 17 Jun 2005 12:04:59 +0000

If you are asking what Apple wants, then they almost certainly want you (and every other developer) to start shipping Universal Binaries regardless of the size. As well as the basic Apple principle of not worrying (or asking the user to worry) about disk space, bandwidth or processer usage right up until its lack becomes a usability issue in itself there is also the fact that Apple will be doing its best to make the dual processors totally transparent to the average end user.

This would suggest that you should automatically offer the Universal Binary to website visitors with Macs and only offer single processor builds (if at all) to those who specifically request them and who can therefore be assumed to know what is going on.

Actually, how much of the download would actually be duplicated in a Universal Binary? I take it all the XUL stuff, images etc. could be shared. Does anyone have the numbers from Apple’s Firefox port?

By: David Illsley

David Illsley — Fri, 17 Jun 2005 08:40:10 +0000

Well I have very little to do at the moment so I downloaded the source to WebKit from http://webkit.opendarwin.org and did a little hunting. I suspect the current state of affairs is summed up by this line:

// FIXME: Some day we will start reporting the actual CPU here instead of hardcoding PPC.

By: Robert

Robert — Fri, 17 Jun 2005 01:52:53 +0000

@Philippe: nobody says you need to force them. Mozilla.org sniffs and recommends a download on the home page, but you can still choose which one you want to download. It just makes it easier for someone who doesn’t know the technical side of things and “just wants what they need”.

By: Philippe

Philippe — Fri, 17 Jun 2005 01:41:06 +0000

Hmm, sniffing for the UA string has never been a very reliable method in my book.

Besides, in your example, what if the user *wants* the universal binary, because he/she will install it in on multiple machines (I downloaded FF 1.0 once, and installed it on 20 or so different Macs).