The illusion of control in web design
alistapart.com
alistapart.com
In Python and C everybody accepts that new versions introduce new features that are not backwards compatible and that errors stops the normal execution flow. For some obscure reason people complain when JS functions stop executing when uncatched errors occur. Or the classic: people complain about surprising result when you add values of different types. They can be happy this produces anything at all! :D
It's a mystery to me why Web Development is treated and evaluated so differently from "normal" Development.
Team works for months getting it down below one second.
App launches, business team loads it up with so much 'analytics' software that page load time is... 4 seconds.
Tear out hair.
It wasn't until I demonstrated what that would be like for a sales rep using a VPN on the road or at a suppliers in India that my boss started to realise why I care about crunching assets, rewriting code to not do stupid things etc.
Speed is nearly always good because of the edge cases.
12 years later, I honestly can't remember visiting the site since despite being in the field and owning a web development shop.
I've googled possibly thousands of web dev questions every year and not once has AListApart come up anywhere in the results.
Amazing to see that they're still around and producing content.
"Two hundred of his employer's sites were instantly reduced to plain old semantic HTML. No CSS. No images. No JavaScript."
I am an end user. I read the text of the webpage. I believe I recieved the full import of the author's message. To read the webpage, I only utilised "plain old semantic HTML. No CSS. No images. No JavaScript."
Questions: 1. Is the page (in the way I the end user consumed it) "robust"? 2. Did I the end user have an "engaging" experience? 3. Who other than I the end user is able to answer questions 1 and 2? How? 4. Do all webpages need to be vectors for the delivery and automatic execution of (e.g. JavaScript) programs in order to be useful to end-users?
Answers: (speaking only for myself) 1. Yes 2. Yes, enough for me to comment 3. Not sure; seems like a subjective determination only the end user can make 4. No. For example, I read the webpage that is the subject of this comment without using graphical fonts, CSS, images or JavaScript. I focused on the author's words alone. I do not believe I missed anything in the author's communication.
The author's single main point (as I read it) is that he just wants his content to "keep working" without having to put in any ongoing effort or maintenance, and without keeping up with relevant news.
That's never going to be a reality in an ecosystem like the web, and I'd argue it's probably not a reality in any consumer ecosystem available today (mobile/desktop/server/etc). They all update, and require applications to keep up with at least minor tweaking.
The part I lose patience with is this: he's advocating for a complete loss of user security and privacy in favor of ease of use for HIM, the developer. Because he won't or can't do his damn job, or because he values their (and by extension: mine) privacy and security so little he's not willing to put in the time or effort.
They are a broker agency for a major financial institution, and all of a sudden, they could not log in to their major vendor's financial transactional system - because it was secured with a GeoTrust cert that didn't have the appropriate root chain and was thus rejected by Chrome. Complete stop. In many ways, much worse than the whole Y2K thing.
The parent financial institution had no idea about the cert, and it wasn't until I sent an email to their IT department that they realised what had happened (because internal tests (with different browsers or older Chrome versions) were still working fine).
It is getting to the stage where we need to have some sort of overriding, holistic approach to managing these sorts of issues and upgrades. It is too easy to think everything is ops normal if you aren't using a particular browser or framework etc. I fully admit that we are guilty of this too - we only test our web development work on a narrow subset of browsers and hope for the best. That target is getting smaller and smaller every day.
There are some benefits to a certain level of heterogeneity.