Ask HN: Why did 4GL tools like FoxPro, Clipper, Clarion, DBase disappear?
These combined database and programming language products were described as 4th Generation Languages (4GL). They were powerful and far more accessible to beginners or casual users than stand-alone programming languages. Why did they disappear?
These 4GL products seem tailor-made for web projects that require database access. Where enormous scale isn't required (i.e. the majority of web development projects), the 4GL approach seems like a perfect fit. Instead we've settled on a mish-mash of clumsily glued-together (and needlessly complicated) solutions for web development with databases.
Why did 4GL fail to adapt to the web era? Were the tools and language too restrictive? (Hard to believe this when CRUD is the bread-and-butter of many web apps. And these 4GL tools made accessing data far more easily and seamlessly than the solutions today.)
Or was there some snobbishness amongst programmers who didn't view these products as 'real' programming languages?
(
10 comments
| 15 points
| on HN
)
Was doing some great stuff with FoxBase/FoxPRO on the Mac. After that mess with dropping of support my next DB platform was open source (LAMP) way more flexibility and am still using it.
The languages were usually some flavor of "business BASIC" which were in reality pretty bad. It was far too easy to end up with "ball of mud" apps. I worked for one company on a project over several years to replace a foxpro app with EJB 3 backed desktop app. A lot of the complexity in that project came from documenting and replacing what users were doing by manually updating the tables with proper workflows (permissions to perform actions? egad!). Note this is not the fault of the 4GL in particular, but more a result of the "throw something together" ethos (so I maintain). So my take away here is: great for prototyping, not so great for long term maintenance and support.
There are still companies offering 4GL solutions in this space like [1] and [2]. I have no doubt there is some enterprise version of Microsoft Access that will expose your app to the web, with an appropriate licensing fee.
Another thought I have is all the X compiled to Javascript to run on the browser languages, like scalajs and java (GWT). GWT (in my view) was a failed experiment, I don't know of anyone using it for new development, all the apps I see are legacy. I welcome feedback on this.
Something else I think about is; in those days, although lisp and smalltalk were around, doesn't seem like they were widely used. Nowadays, there's an abundance of dynamically typed, highly productive languages with extensive library ecosystems that essentially enable developers to move the language to meet (any particular) problem domain, and create solutions quickly. So the appeal of environments that bind language/database/display together isn't as novel as it once was. In other words, Ruby and Rails is your 4GL today.
> Or was there some snobbishness amongst programmers who didn't view these products as 'real' programming languages?
Definitely some snobbishness; but also the fact that as developers, we have to chase 'the new hot' on some level to stay current, and marketable. No one wants to be stuck only knowing coldfusion, when those jobs all go away.
[1] https://www.elevatesoft.com/products?category=ewb
[2] https://harbour.github.io/
I need to get around to try eve. So far I got by using Jupyter/iPython notebooks for similar purposes, which means I don't have as much of a need.
If you tried eve already, what did you build with it?