3 ms·
>Web developers have utterly squandered all efficiency gains of the last 30 years It's true, whenever a company has pleaded with me to bring a site in with bet
by bryanrasmussen 6y ago
>Web developers have utterly squandered all efficiency gains of the last 30 years
It's true, whenever a company has pleaded with me to bring a site in with better performance, I have adamantly refused to do so.
When companies say guess what, Bryan, we are going to focus a sprint on just making sure everything download as fast as possible and we get rid of anything getting in the way of the best possible user experience, I have spent that sprint watching quirky animation, and sometimes turned that animation into a base64 encoded gif and put it in a div that was set with a z-index low enough that it would be not seen by people but would still have to be downloaded by the browser!
I do all these things of course, because it was decided by the Secret League of Obnoxious Web Developers (SLOWD) that we should do the utmost in our power to make the web slower for everyone.
OR - it could be that I have in fact asked project management at sites repeatedly to focus on performance (and accessibility, another thing that always gets ignored) and been told that nobody care about that stuff.
I guess it's up to you to determine where the fault lies.
- stonogo 6y agoI don't think anyone is under the impression that web developers can't make excuses; we have plenty of evidence of that. I think discussions like this are capable of serving as valuable reference material when we are engaged with project managers on this topic.
- bryanrasmussen 6y agoI don't think there is a paucity of research that shows that higher performing sites get more users and are generally more valuable for the site owners. The problem is not lack of reference material, the problem is that nobody considers the problem that important (generally) when presented with that reference material. Now the thread is quite big so I have not read everything but I have skimmed through it all, and have not seen anything that would be a really useful argument for getting someone to consider maybe we should try to improve site performance. A bunch of people complaining about Google, Wired, and a few other things that they complain about is not as impressive as just the hits you get if you search "effects of site performance on user retention" (replace user retention with other useful metrics to get sources for the argument being written up)
- thih9 6y agoLooks like you’re saying that developers can’t discuss priorities with management. In some companies this might be true. But in many, when a CTO or a senior dev demonstrates business value of a speedy website, it would get attention. I’m sure project management doesn’t care about keeping packages up to date either. But most devs can successfully communicate that this is necessary and that the alternative is way worse.
- bryanrasmussen 6y agoYes, when I was CTO I did a lot to make our website speedy. When I was consulting I would often do performance analysis of the sites I consulted at, show how performance improvements could be made, put links to relevant studies on performance improvements and the effects on the bottom line with nice quotes, to have the task of cleaning everything up be put in the backlog and forgotten forever. >I’m sure project management doesn’t care about keeping packages up to date either. How many days or weeks does it take you to update a package? It generally takes me a minute, sometimes problems happen and I need to spend some hours but those are infrequent. If a package update is going to take too long it becomes something for project management to be aware of and sometimes it is not allowed to update a package. But generally issues with site performance need handling over a longer period of time than updating a package, I would think that was evident to anyone that has ever updated a package or done a performance analysis of a site. The comparison between something that generally takes a minute and something that takes weeks seems insincere. But I can make an example where an update was needed that would take a significant amount of time, the reason that the update was accepted was that it would fix certain bugs with the old package and it needed to be done. Either the package was updated or the bugs were allowed to remain or we could fix the bugs with longer time than it took to update the package. I think it is obvious how this is a different argument than the site performance thing, the package update is an argument that this way we will fix the problems that you the business have pointed out, the site performance thing is an argument that this is a problem we want you the business to acknowledge.
- tjpnz 6y agoI once worked for a place where the business unit got us to deploy one of those crappy client side A/B testing scripts which blocks rendering. Conversion rates started to dip and the same people came back to us complaining about it. I was able to pull together plenty of evidence to show them what the issue was. All they cared about was if we could A/B test it using the aforementioned client side script. Some people just can't be reasoned with, so glad I got out of that place.
- greggman3 6y agoI prefer fast non bloated websites but your example does remind me of a tangential example. Video Game Load times. People complain but they really only complain if the content is not great. If the content is great no complaints. Case in point, any game by Valve has atrocious load times, even and including their latest, Half Life: Alyx. And yet Valve's titles are among the highest rated games. And so, load time is rarely prioritized because it clearly doesn't matter. What appears to matter is content people want. (T_T)
- bryanrasmussen 6y agoI think the context of what you're serving matters. In this case the game content is good enough that the performance probably doesn't matter. The worst example of a refused performance improvement I can think of was in relation to improving a help site, it had generally bad reviews from users (it is very difficult to get good reviews from users for a help site because if a user is coming to your help site they are already in a bad mood) but obviously if you are on a site that you are mad about being on and it then takes a long time to load all the data so you can try to figure out your problem you are going to be steaming. Project manager wouldn't prioritize the three performance improvement tickets I made with lots of cogent description of why it needed to be done. Somehow though, this is my fault.