Average Web Page Size Triples Since 2003
[found via a whole lot of sources I guess, Techmeme, Slashdot, and so on, original at link ] While I guess it is fair to say I would not be known for generally having a strong focus in the area of UI/Page side of things (oppps hope the current crew is not watching ;) ). I found this article on the Average Web page size tripling on the top 1000 Web sites to be very interesting, as I have in the past spent quite a bit of time in this space in the distance past (or it feels like the distance past now :) ) The report touches a lot of interesting points about Web2.0, trends in Web pages and all of it is good stuff, but I think it compressed a lot of important aspects that need to be highlighted - that perhaps of all places is best highlighted by a post on valleywag, "Ignoring customers still using modems is officially all the rage. Usability and accessibility guru Jakob Nielsen to shake fist at you all." I really like the fact that a report like this has appeared, as I think it highlights very succinctly the technical/transmission realities of how Web2.0/et.al. is having on end user experiences both in a positive and negative sense. To even emphasise this point, I believe that in the valley a lot of companies like Yahoo, are getting folks to focus on the "transmission" aspects of a UI/UX and in the case of Yahoo have led to the birth of Y!Slow etc... There is some video's on the Yahoo! developer network from the Director responsible for this team, sorry I don't have a link :( Off the top of my head, I reckon what would have been nice to have in the report would have been the following: - use of sub domains
- the issue of location and speed of light, i.e. the value proposition of a Global CDN
- HTTP protocol versions, pipelining, browser versions/differences and so on...
- CSS and JavaScript placement in HTML etc....
- and many of the recommendations that get highlighted from Y!Slow as I recall.
- (plus I would have a few, which I will mention below and in a follow up post)
I guess overall, one of my initial thoughts and critiques of the article is that it could have given some link love to the all too few links, presentations and tools that help alleviate many of the concerns the article raises about the "un-optimised" transmission costs of an ever increasing number of embedded objects in a Web page that ultimately better than we have in the past (on average). Yeah I could link to YSlow, but that would not be no fun :) YSlow is the best and true place to start, as what I am thinking about blogging about as a series is at a different level and builds on YSlow's baked in guidance. So what would I further add to the conversation!!! Well I guess a rough "re-call" of my experience in the technical/transmission issues of delivering superior UX at the point of delivery that matters, i.e. the area of perception. A hard/real example of this, is why from a transmission point of view does Google's home page work, - simple page (not applicable to many/all scenarios, but something to bear in mind)
- Page is compressed when possible, and have few/little embedded objects
- redirects to shortest "transmission path to delivery", i.e. date centre next door is king in almost all cases
- (ok simple so far, right!!!! But GOOG homepage is a jump point and not an "aIttention sink" so lets keep going)
So....... I better state that I am stating from recollection and obviously there is not an ounce of warranty on where I believe (or understand) this current space is at (or has transitioned to), as I have not tracked it in years, but i thought it would be interesting for those that might have commercial interests in demographic patterns that perhaps present geographical/regional/adoption challenges. In a word, I am talking about PEP's :) These are a Web1.0 (and pre-) concept that I personally have had more than a healthy historical vested interest in, albeit today I am very rusty, but since the "Internet" circle of life is doing the full circle, why not :) So yeah I am *REALLY" talking about where the rubber hits the road! Today's world is based on so many "old" design assumptions and while 20-25+ year old design assumptions on base protocols are really the key-stone of the success of the Internet, my key observation is that the ubiquitous nature of the web will bring us all the way back to the OSI layer. Sounds weird and certainly not for all, like many if not all of the Web 2.0 optimizations, but when it does matter it really matters a lot. For example, I would argue that there is too serious of a focus on "the happy" case when it comes to "dynamic" content right now! Why? Besides my pseudo (and poor mans attempt/argument for the following cases/) argument for what is assembled as as an always-on PC/Desktop/MacPro based device and with some of the following - an optimised "page", assumed, with YSlow like recommendations etc....
- Global DNS - assumed?
- (The Elephant in the Room, for some, not many, but some, and in different ways)
and then you have ... a whole world of future costs... Non ("US") Web 2.0 space etc....... :) which I will cover /provision thoughts in the next post :) /Paul ps Have no doubt I have mistakes, etc.... in this post, chime in as you see fit, this is the aim.
Web 2.0
Monday, April 28, 2008 9:31:49 PM UTC |  | Disclaimer | |
|
|