In addition there are standard ways of encrypting/signing XML documents.
I still don't want to go back to that time, mind you. Is probably why I'm so hesitant on the current attempts.
- Go and similar "one binary". Bonus points if static files are also embedded
- "Just a JAR with few parameters"
- Anything that comes pre-packaged in OS. Perl with OS-provided libs.
- Enterprise Java with code-as-xml for config. More pleasant if we're not the ones managing the XMLS
- C/C++ apps with reasonable deps that don't need re-compiling half of OS
- Bare containers. Shit visibility without a lot of tooling but at least we don't have fight deps
- k8s. Complexity of system rarely pays off with the relatively simple apps our devs are deploying and need even more tooling for good visibility
- Ruby outside of container - just fucking kill me. A bunch of requirements for libs in system just to compile "gems" (why those cunts can't call it libs like everyone else), long startup, leaks memory, took them 10+ years to make a web server that's not completely shit...
- Python - if I see something is in Python I try to find alternative solutions. I swear half of the install instructions either just don't work or shit all over home/system dirs. The absolute clusterfuck of py2-3 migration didn't help either, thankfully it appears its finally over. Bonus demerit if instructions mention setup.py. Tho to be fair I didn't tried recently, maybe it is better
Most vocal java haters I had met were JavaScript programmers who knew basically nothing about Java. Their hate was purely cultural - they learned on discussion forums that they should hate java. But, they had literally zero knowledge whatsoever.
oh, I got one to admit he was "intimidated" by Java.
Note: there is a subset of developers who have either hit one of Java/JVMs limitations, or, subscribe as a matter of professional judgement and taste to a school of programming (say FP) that is not a good fit to Java and JVM's orientation (classes, OOP). This subset is an informed subset and their feedback is reasonable and at times has helped Java/JVM move forward.
Code as json or yaml is fundamentally the same while being nicer to look at (yaml at least).
So you need to navigate pitfalls of YAML (or bear ugliness of JSON) all while learning a templating language.
All because someone decided "Users will not want to learn Python for DSL, let's give them YAML"...
Every single tool that went that route has that problem and every single one would be better served just picking <language> (Lua is nice and embeddable, just saying) and using it as DSL.
For instance, in Quarkus, you can have a full REST application in about 12MB of RAM and running on the Graalvm.
In Spring (often looked at a heavy framework), this is the code required to write a full REST endpoint:
@SpringBootApplication @RestController public class DemoApplication {
@GetMapping("/helloworld") public String hello() { return"Hello World!"; } }
IDK, I think the irony is you thinking IDEs wouldn't use those patterns. I bet you'll find tons of weird interfaces and abstractions in IntelliJ. It is a huge and complicated code base.
The IDE that's written in Java most likely shares nothing with Java ecosystem parts they "hate" (I'm guessing all of the EJB, applets an general 90's Java enterprise stuff). Only irony here is you thinking it is.
The 90's Java was a thing to hate. Running fat complex app server with bunch of code-as-XML for configuration just to run simple API app. IDEs are not doing that (althought Eclipse was a bit of a monster of glued in parts)
> Did anyone in the mid to late nineties / early 2000s really discover Java and Java applets and thought: "Java applets is the tech that'll catch on, I'll invest in that" and not invest in Java itself?
...entirety of management it seems, it was the most popular hammer for the problem for quite some time in enterprise. Hell, we still have some less than decade old servers needing java applets for KVM