I remember converting the HTML on my website to XHTML... and then a few years later having to revert it to HTML. What always bugged me about designing XML was trying to decide what should be an attribute versus an element. And while I loved the *idea* of XML, I always found it visually noisy. These days I prefer JSON.
I remember when XML was initially released or at least available for use in browser code and was only supported in certain browsers. For our applications we needed to force users to use IE so we wouldn't have to write 5 versions of our code. XML enabled us to pass data in a structured way from browser to server and it was awesome. Most of our users were on dial-up so anything that enabled us to keep another page from loading to do data entry or retrieve data was a godsend. ... years later JSON enabled more standard utilization of application code on front-ends and as broadband became more ubiquitous, heavy pages became less of an issue. Progress is fun and keeps developers employed rewriting.
Wonderful article! A bit surprised that Agile got left out. Gotta be one of the biggest hypes in the business during the last 25 years! 😁 (Not saying nothing about it is useful, just that there has been a LOT of hype!)
I disagree with your comments about XML, but I agree with most of the rest of it, starting with XSLT.
Some of these technologies are good and appropriate, for some applications. There was NoSQL long before there was SQL. There are *disadvantages* to doing everything in SQL. It is *NOT* a perfect solution to *all problems in the world*.
But yea; there's a lot of hype and trend-chasing.
When I first learned of XML, it was *Hyped Like Mad* by nearly everyone!!!
My response was, "This is obviously good and useful. But would you please relax and calm down?" I said, "When the hype is over, we'll find that we're using XML for everything without even talking about it. You'll say 'I'll send some data.' and people will assume XML." And yes, that has happened.
(OK; JSON, YAML, and a couple others have taken off. JSON is arguably "simplified XML." And YAML is simplified and more readable.)
Thanks for the post. This made me recall an older talk of Hadi Hariri, titled "The Silver Bullet Syndrome". It is an old (according to our technologies lifecycles) talk but still timeless, indeed copying is not engineering & following hypes and being at cutting edge is not architecting.
This is what happens when you judge people by the complexity of the systems they create, not by the complexity of the problems they solve. Then again, if we did that, this entire industry might just completely collapse.
Such posts often make it feel as if there's no real applications for all that tech in the browser. There is, that's how we get Figma, Spotify, Notion, Photopea or any other app. Still, it's true that the way people think about web apps has changed and building a very basic app requires a ton of things ("wtf is webpack?") compared to the Apache+PHP+MySQL era.
Thank you for this, I absolutely loved it. Luckily I have seen things similarly and stopped at using XML and XSLT for certain things that it was/is good for, but otherwise I just stuck to what was working for me in the first place. This basically made me unemployable because if I heard somebody have these you mentioned as requirements, I looked at them with "are you a f* idiot" clearly readable from my facial expressions. Let's hope people will actually find their way back to normalcy.
Yeah. The problem isn't that this tech isn't useful, it's that when someone like Google is using something, _you_ probably don't have any need for it. Serverless has applications. So does gRPC. So do React and TypeScript and AMP and all these other things. If you're building an MVP to see if there's even a market for your product, and you're getting bogged down in Helm charts and Elastic MapReduce, that's a _you_ problem, because you should've picked the quickest thing to get you to market. Hype-based technology choices will always be a bad thing, whether it's transitioning to on-prem and vanilla JS or doing K8s for a marketing blog. (Just to be clear, I'm agreeing with the author, just summarizing in slightly more aggressive terms.)
I remember converting the HTML on my website to XHTML... and then a few years later having to revert it to HTML. What always bugged me about designing XML was trying to decide what should be an attribute versus an element. And while I loved the *idea* of XML, I always found it visually noisy. These days I prefer JSON.
Excellent article as usual :)
And yes, I like the new "hype" about YAGNI, No serverless / micro-services anywhere, HTMX / alpine.js :p
This was written by my spirit animal. While I recently wrote a rant about microservices, this covers way more ground in a more succinct way.
Good Article.
The XML/SOAP part reminds me of https://youtu.be/RnqAXuLZlaE?si=lFYnuDNEQJvTOfxW
Same subject, 20+ years ago.
https://www.joelonsoftware.com/2002/01/06/fire-and-motion/
I remember when XML was initially released or at least available for use in browser code and was only supported in certain browsers. For our applications we needed to force users to use IE so we wouldn't have to write 5 versions of our code. XML enabled us to pass data in a structured way from browser to server and it was awesome. Most of our users were on dial-up so anything that enabled us to keep another page from loading to do data entry or retrieve data was a godsend. ... years later JSON enabled more standard utilization of application code on front-ends and as broadband became more ubiquitous, heavy pages became less of an issue. Progress is fun and keeps developers employed rewriting.
Right!
Wonderful article! A bit surprised that Agile got left out. Gotta be one of the biggest hypes in the business during the last 25 years! 😁 (Not saying nothing about it is useful, just that there has been a LOT of hype!)
And scrum, XP, etc. I use them a lot, but yeah the hype was so over the top, and people that were hyping the most were often doing it all wrong.
I disagree with your comments about XML, but I agree with most of the rest of it, starting with XSLT.
Some of these technologies are good and appropriate, for some applications. There was NoSQL long before there was SQL. There are *disadvantages* to doing everything in SQL. It is *NOT* a perfect solution to *all problems in the world*.
But yea; there's a lot of hype and trend-chasing.
When I first learned of XML, it was *Hyped Like Mad* by nearly everyone!!!
My response was, "This is obviously good and useful. But would you please relax and calm down?" I said, "When the hype is over, we'll find that we're using XML for everything without even talking about it. You'll say 'I'll send some data.' and people will assume XML." And yes, that has happened.
(OK; JSON, YAML, and a couple others have taken off. JSON is arguably "simplified XML." And YAML is simplified and more readable.)
Thanks for the post. This made me recall an older talk of Hadi Hariri, titled "The Silver Bullet Syndrome". It is an old (according to our technologies lifecycles) talk but still timeless, indeed copying is not engineering & following hypes and being at cutting edge is not architecting.
For those who wonder: https://www.youtube.com/watch?v=3wyd6J3yjcs
This is what happens when you judge people by the complexity of the systems they create, not by the complexity of the problems they solve. Then again, if we did that, this entire industry might just completely collapse.
Such posts often make it feel as if there's no real applications for all that tech in the browser. There is, that's how we get Figma, Spotify, Notion, Photopea or any other app. Still, it's true that the way people think about web apps has changed and building a very basic app requires a ton of things ("wtf is webpack?") compared to the Apache+PHP+MySQL era.
Figma is mentioned in the article, and the conclusion of the post support s conclusions.
En comparaison avec le français, le javascript est parfaitement logique.
How can a software project fail in mass? Too much, or too little?
If you analyze the subtext carefully, you will understand there is a deep meaning behind it, which is...
... that the author is French and badly translated the words in his mind.
And he is going to fix that now.
The words, not being French. There is no cure for that.
Thank you for this, I absolutely loved it. Luckily I have seen things similarly and stopped at using XML and XSLT for certain things that it was/is good for, but otherwise I just stuck to what was working for me in the first place. This basically made me unemployable because if I heard somebody have these you mentioned as requirements, I looked at them with "are you a f* idiot" clearly readable from my facial expressions. Let's hope people will actually find their way back to normalcy.
Yeah. The problem isn't that this tech isn't useful, it's that when someone like Google is using something, _you_ probably don't have any need for it. Serverless has applications. So does gRPC. So do React and TypeScript and AMP and all these other things. If you're building an MVP to see if there's even a market for your product, and you're getting bogged down in Helm charts and Elastic MapReduce, that's a _you_ problem, because you should've picked the quickest thing to get you to market. Hype-based technology choices will always be a bad thing, whether it's transitioning to on-prem and vanilla JS or doing K8s for a marketing blog. (Just to be clear, I'm agreeing with the author, just summarizing in slightly more aggressive terms.)