3 ms·

>Enterprise software sucks because it is a hard, extremely complex problem space. I couldn't disagree more. I don't know how many enterprise systems I've worke

by jkush 19y ago

>Enterprise software sucks because it is a hard, extremely complex problem space. I couldn't disagree more. I don't know how many enterprise systems I've worked on that are unnecessarily complex.

DanielBMarkham 19y ago
Solving multiple, overlapping problems for thousands of users all over the world is non-trivial. Pick any solution space, and I can give you a dozen things you haven't thought about. 95% of them aren't important to one or another subset of users, and probably actually get in the way of their work. The problem is that the 95% of feature bloat isn't the same for all the users! The trick is not to deny complexity or embrace it. The trick is to make rational compromises. Start with something simple and add just enough bells and whistles to get most of what's needed done. But that's a much more painful and difficult process than just buying some big product off the shelf.
- jkush 19y ago
  Great points. What I mean by unnecessarily complex is: I've seen to many enterprise systems that are simply electronic versions of manual processes. The problems that enterprise software system really try to solve are masked by trying to genericize them for every situation and everyone. It still comes down to solving one specific problem and solving it well.
  - DanielBMarkham 19y ago
    We are very close to agreement. There's hard work in picking the best and most competitive business processes, and it's going to be different for every organization. That's why "project management software" isn't the same as "software to help us plan and control building and deploying industrial power transformers in a union environment where we have high governmental reporting obligations" or some such. Companies either want to sell generic products with feature bloat all over the place, or home-grown shops try to take fragile systems and glue them together with more and more code. Either way you end up with a lot more complexity than the organization needs. But it still might not be a simple answer. And nobody wants to do the work of figuring out what the minimum amount of complexity is for each situation. (Even though there's probably more money in that than anything else)

by jkush 19y ago

>Enterprise software sucks because it is a hard, extremely complex problem space. I couldn't disagree more. I don't know how many enterprise systems I've worked on that are unnecessarily complex.

DanielBMarkham 19y ago

Solving multiple, overlapping problems for thousands of users all over the world is non-trivial. Pick any solution space, and I can give you a dozen things you haven't thought about. 95% of them aren't important to one or another subset of users, and probably actually get in the way of their work. The problem is that the 95% of feature bloat isn't the same for all the users! The trick is not to deny complexity or embrace it. The trick is to make rational compromises. Start with something simple and add just enough bells and whistles to get most of what's needed done. But that's a much more painful and difficult process than just buying some big product off the shelf.

jkush 19y ago
Great points. What I mean by unnecessarily complex is: I've seen to many enterprise systems that are simply electronic versions of manual processes. The problems that enterprise software system really try to solve are masked by trying to genericize them for every situation and everyone. It still comes down to solving one specific problem and solving it well.
- DanielBMarkham 19y ago
  We are very close to agreement. There's hard work in picking the best and most competitive business processes, and it's going to be different for every organization. That's why "project management software" isn't the same as "software to help us plan and control building and deploying industrial power transformers in a union environment where we have high governmental reporting obligations" or some such. Companies either want to sell generic products with feature bloat all over the place, or home-grown shops try to take fragile systems and glue them together with more and more code. Either way you end up with a lot more complexity than the organization needs. But it still might not be a simple answer. And nobody wants to do the work of figuring out what the minimum amount of complexity is for each situation. (Even though there's probably more money in that than anything else)