[bena]

Do you think this might be a function of the same thing the electronic medical record (EMR) field faces?

For reference, I used to work in the EMR space for a "value-added reseller" of a certain EMR package. It wasn't great. But when you looked at the other options, they weren't great either.

Eventually I learned that medical software is a ghetto because those who care about the medical part of it don't really care about the software part of it. At least not enough to treat it as the craft it is. And those who care about the software development tend to leave because you can do better work elsewhere.

So do you think GDS products suffer because no one cares about actually developing software, but rather the end goal?

[PurpleRamen]

GDS suck because of compability. Compability to decade old systems and workers, and compability to evertything and anything. The first GDS where 40-50 years ago created, under the stress of very limited resources, aimed for very specific handling. In that time, workflows and concepts were established and brought into productions, which are still used today.

If you think unix-commandline is cryptic in it's usage, then you should try gds-commands, which are often optimized to the single characters. Today you can use XML and JSON, but they are only build on top of those old systems, leading to many old crafty things.