Hacker News Reader

Comment by cygwin98 | Hacker News Reader

I see utter incompetency in the management lack of planning and leadership. Though the hacker inside me find it a good opportunity to practice compiler theories, say, write a compiler to translate modern languages to Cobol, just like CoffeeScript to Javascript.

View on HN

Wouldn't the opposite solution be better? Write a compiler from Cobol to another language? Perhaps C, because C can be called from any other language and can do gotos, etc to emulate legacy control flow and other oddities.

After you have your legacy Cobol code running in a modern environment and the frontend can interface with it, you can start gutting it piece by piece to get rid of the legacy codebase.

I fear that in 50 years, there will be a tremendous need for programmers that can deal with a horrible 1990's language called Java and fucked up non-paradigm called object oriented programming.

I also see this as a problem with the companies who still want to maintain code from decades ago. It will become more expensive day by day. Of course, to a banker, computer programming and it maintenance is nothing but a cost. However, at some point it will become more affordable in the long term to just scrap the old legacy systems.

In MBA school they are taught that you are a cog that can be replaced by a similar one. They have no idea of the weak position they are in. The old COBOL guys have them over a barrel and the system vendors can charge whatever they wish for maintenance.

Since I'm taking a Cobol class this next semester for a similar reason, I went searching for whats out there and found some people attempting that: https://github.com/dayashi/cobor

Reply on news.ycombinator.com