7 Comments
User's avatar
Matthew L Johnson's avatar

As someone the builds the in-house tools: you get what you get when you buy off the shelf. There is a time and a place for those products. But when you want something that truly fits your business style, there is nothing better than a custom designed solution.

Kenny De Metter's avatar

Because those third party tools aren't necessarily better either. In the best case it's going to be good at one specific thing, but that's usually not what the company wants/needs.

Having worked for a large company for nearly 20 years, I have seen the same happen over and over again:

We started with tons of in house developed software, that became harder to maintain and bloated because of a lot of feature requests from users ("Business")

Business loses faith in inhouse development and starts looking for third-party software and cuts developments teams.

They pay lots of consultants for many years who promise them everything will work. The consultants write custom code to make it work which is often more bloated than what the inhouse team initially had.

Business eventually loses faith in the consultants and starts building up in house development again. In house devs builds something that works well initially, but then becomes bloated over time due to many feature requests from the users.

Rinse, repeat.

The problem isn't in house development or third party development. It's that what the user wants ( and wants to pay for) isn't what the user needs. And the user isn't interested in learning why.

Marko's avatar

Also, in house built tools are often custom tailored to one's specific needs. You don't need to depend on someone else promissing they are going to build what you have requested years ago.

Bill de Haan's avatar

There are absolutely cases where the industry standard tools won't do what's needed, and can't be extended or modified to do what's needed, and an in-house tool HAS to be built.

However, a huge number of in-house tools are built because of "not invented here" syndrome. NIH not only has the benefit of employing software architects, it allows them to replace standard tools with "better", and incompatible, in-house tools.

I know companies where the company's architects disliked XAML, so they wrote their own. Of course, WPF wouldn't work with it, so they wrote their own implementation of that, too.

One company had its' own replacements for, or "improvements" of:

- XAML

- SVG

- WPF

- boost libraries

- C++ memory management, and therefore

- C++ pointers

- UMLDOX

- make

Of course, this meant that the company couldn't leverage knowledge of new hires who knew the industry tools, they had to learn the in-house tools (IHT). It often took new hires two years to get up to speed. Improvements in the commercial tools meant that the IHTs written years ago were falling behind, in terms of relative functionality. Some projects costed out development assuming standard features available in the industry that the IHT lacked, and took major losses. Code written in departments that used the IHTs was incompatible with departments that didn't. And when the architect of the IHT left the company, there was usually no one to maintain it.

There are reasons to buy, and reasons to build. But when the argument for building is because the architect says that the IHT will be more "elegant", buy.

As I used to tell management, "brute force has an elegance all its' own"

Bruce Mardle's avatar

... or, as they would probably say at Microsoft, "so we can lock our customers into using *our* version" :-(

Flemming Nørnberg Larsen's avatar

Not Invented Here (NIH) 👍