That wasn't a mandate. It was my proposal so I know. I suggested that we target "about the size of an MP3" because that was the most common unit of download that a lot of people understood in the early 2000s and downloading software was something people weren't terribly comfortable with. I did some quick estimates and came up with 4MB as the goal based on the size of popular MP3s. By the time we shipped Firefox 1.0 in late 2004, we managed to get the download to about 4.7MB.
That was 2004 though. Today the Web is a lot more capable because browsers are a lot more capable. Also, typical download bandwidth for consumers is a lot more capable than it was in 2004.
Most of the growth in Firefox download size since then is a result of Gecko/Web platform feature growth, not the GUI features that users interact with.
That web platform capability isn't free. It takes code to make JS dozens of times faster than it was in 2004. It takes code to add HTML5 and CSS 3 features, WebGL, WebRTC, and all of the other great stuff the Web platform includes today. That code makes the download larger.
That being said, I'd love to see another round of evaluation to see what can be trimmed or slimmed. I don't consider that the same priority it was in 2004 though.
That was 2004 though. Today the Web is a lot more capable because browsers are a lot more capable. Also, typical download bandwidth for consumers is a lot more capable than it was in 2004.
Most of the growth in Firefox download size since then is a result of Gecko/Web platform feature growth, not the GUI features that users interact with.
That web platform capability isn't free. It takes code to make JS dozens of times faster than it was in 2004. It takes code to add HTML5 and CSS 3 features, WebGL, WebRTC, and all of the other great stuff the Web platform includes today. That code makes the download larger.
That being said, I'd love to see another round of evaluation to see what can be trimmed or slimmed. I don't consider that the same priority it was in 2004 though.