Website Optimization Measures, Part I
Jens Meiert, February 10, 2008 / May 5, 2008.
This entry is filed under Web Development, User Experience.
Permanent focus on QA includes more thorough website revisions from time to time, and that does certainly not mean a “redesign” or “relaunch”. This week I spent a lot of time analyzing, refactoring, and optimizing some of my private sites, and I thought I’d quickly share a few of these things for inspiration and discussion:
-
Removing irrelevant posts. I indeed browsed this blog for irrelevant posts – and I found a lot. While some might generally consider deleting posts “taboo”, I do not, as the posts in question decreased the average post quality. I accept that not every post needs to make it #1 on Digg, but admittedly, some entries looked like I just babbled in order to say something. And while this removal measure also makes maintainance a little bit easier (apart from the fact that I need to remove some redirects later), it didn’t affect many comments – few people commented those posts, of course, probably. So actually, there’s no downside with removing irrelevant stuff, not even in blogs.
-
Editing and removing comments. Some might consider this yet another taboo, but it has a few very good reasons, partially mentioned in the new comment guidelines. The most important reason was spam, and my tolerant approach to comments that avoids both moderation,
nofollow, and stuff just bears a larger risk of getting spammed badly without noticing it (even though I do check everything Akismet does and what gets published). So I went through all comments, removed some, edited or “neutralized” others. While this measure meant a lot of work it increased this blog’s overall quality and decreased the slim likelihood of getting “penalized” anyway – and it benefits all commenters as well, by emphasizing and strengthening their contribution. -
Cleaning up the folder structure for “auxiliary” files. For style sheets, scripts, images, and stuff I personally used to use a folder structure like “/bin/css”, “/bin/js”, and “/img”. Not anymore. The mentioned structure has been inspired by UNIX stuff anyway but doesn’t make real sense, and it doesn’t scale very well either. So this architecture has been changed into
- /media
- /setup
- /css
- /js
And while this appears to be much better than the former type of organization, I’ll observe how it really behaves – so far, it works and scales great. “/media” is not just for images and thus gives me the flexibility to throw in videos as well (something I didn’t really do before, not in my private projects), and “/setup” looked like a short name that would roughly legitimate hosting style sheets and the like. Anyway, if there’s something better, I’ll revise it again. Unless cost of problem gets even lower, that is.
-
Revising use of
rel="alternate". I used to addlink rel="alternate"elements in theheadsection of some documents to point to alternative versions (mostly translations) of the corresponding page. I dropped this, not due to eventual SEO advantages but due to maintainability reasons. Addingrel="alternate"only to hyperlinks pointing to alternative versions now needs to suffice, and it will. -
Replacing Google and Yahoo verification
metaelements by HTML equivalents. Using both Google Webmaster Tools and Yahoo Site Explorer I once decided to use themetaelement verification way in order to avoid additional stuff in the corresponding project roots. However, I “changed my mind” since this obviously has a downside, namely forcing homepage visitors to download these useless elements, useless for them. By removing themetaelements and verifying sites with the alternative HTML documents file size decreased again, and that is great.
Well, since leaving Bremen tomorrow for a few days of home search in Zurich and for the next gig of my farewell tour in Berlin, I needed to hurry a little bit and kept some arguments and explanations a little bit short. But, at least four additional measures are waiting in part II, and I had a less pleasant feeling with several other English posts anyhow. See the first refactoring measure, right 
This has been the first part of an open article series. Currently there are three additional articles on website optimization, part II, part III, and part IV.
Read More
Enjoy the most popular posts, probably including:
Comments
-
On February 12, 16:47 CET, Stefan Nitzsche said:
From my point of view, it makes sense to think about a blog post before you publish it, and not with hindsight. I think that users don’t like to comment on posts that will be possibly edited or deleted during a “blog review“. And much more important: I would never touch a comment, because the writer meant it the way he wrote it. My changes would only affect the typographical, grammatical and orthographical correctness.
Live with your past babble and look forward to make it better.
-
On February 14, 22:41 CET, Jens Meiert said:
Thank you, Stefan, I guess you point out a few things other readers might worry about as well:
From my point of view, it makes sense to think about a blog post before you publish it, and not with hindsight.
Of course. Alas, mistakes occur, and to quote Confucius:
A man who has committed a mistake and doesn’t correct it, is committing another mistake.
I think that users don’t like to comment on posts that will be possibly edited or deleted during a “blog review“.
Maybe I didn’t make myself clear about that: Unless comments are “spammy” (or violate the – from my point of view – reasonable comment guidelines), no comment gets edited or removed. I absolutely understand concerns about this measure, but it is supposed to benefit all people involved. (And/though it will and can not happen that often, either.)
And much more important: I would never touch a comment, because the writer meant it the way he wrote it.
Neither spam nor insults? (The latter didn’t happen in my blog yet, but I wouldn’t tolerate insults in any way.)
My changes would only affect the typographical, grammatical and orthographical correctness.
Great thing with posts, “questionable” thing with comments, but presumably, you didn’t refer to the latter.
