How much optimization is REALLY needed

When is short TOO short

[Marshall]

Optimize the HTML. Optimize the CSS. Optimize this and that. Okay, so where does one draw the line? When is code too short and optimized too much? So you got your CSS from 7 lines to 3 – big deal. Is that really going to impact usability/viewability? (I know, not a real word) Is the nanosecond saved on the download time truly worth the amount of time you put into it? And considering the increase in hi-speed access, does it matter. In the end, the one item that really slows down a page are images.

Bottom line – when does optimization become a waste of time?

Marshall

[zCat]

Bottom line – when does optimization become a waste of time?

When it starts impacting your own productivity by making maintenance difficult.

The only people who really have to think about maximal optimization are the really big sites, where over millions of page view, the odd few bytes here and there can really add up.

[Kurgano]

As the results suggest... first place, second place, third place... it's a race. If it were a car race you'd want to tweek until you could compete consistantly but you wouldn't stop and neither would the others, its a never ending process.

In early 2008 the new php version is slated to be released, it has already been warned that the new version may break php4 code. Webmasters, many of them, will re-write their code to take advantage of whatever new features become available. It's truly never ending.

The real bottom line is that without knowing what the standard is (algorythms are top secret) theres no way to know an answer to your question. If the algorythms were public - honest webmasters could ensure they comply and be done. Unfortunately dishonest ones would find ways to manipulate the algo ruining it for all. There is no answer to your questions besides "if you're happy with where your site is at, you're done."

[lavazza]

When does optimization become a waste of time?

When the costs outweigh the benefits, for the shareholders and the stakeholders

One way of assessing the wants and needs of the 'key stakeholders' is using 'MoSCoW rules':
---------------
Must Haves - elements/features that are essential to the success of a site
o
Should Haves - important but success is not reliant on them
Could Haves - can be omitted without adverse impact
o
Won't Have - will be left out (for now), might be done later
---------------
Of course, these 'rules' are completely subjective, as are the terms 'success' and 'key stakeholders '

In the end, the one item that really slows down a page are images

A picture is worth a thousand words
1000 words < 10kB

If byte-size is the yard-stick, then 'code optimisation' is a waste of time (i.e. money)

Rather than spending 2 hours reducing seven lines of CSS to three, I'd spend 2 minutes adding another 10 lines in /* comments */ to explain what it is that the code is doing

When cross-platform accessibility and code maintenance are considered to be Must Haves or Should Haves, this type of 'optimisation' is almost certain to pay dividends