4 ms·

Those accessible programming environments tended to exhibit exponentially worsening performance and maintainability characteristics as the complexity of the sys

Those accessible programming environments tended to exhibit exponentially worsening performance and maintainability characteristics as the complexity of the system being built with them grew. This would be have been fine, so long as these systems remained relatively simple, and were rebuilt in more heavyweight but scalable languages once they reached a certain level of complexity. But that isn't what happened. Instead, they just kept growing into twisted, unmaintainable monsters that organisations were completely reliant upon, but which were an absolute nightmare to maintain and extend. As a result, the industry shifted to encouraging the use of "real" programming languages for everything, even the simplest systems, because the greater up-front costs were saved many times over by having a sane story for long-term maintenance and improvement.

thrower123 6y ago
Is this really the case, or does stuff just continue flowing into the tools that are still accessible to users? I have seen an incredible number of businesses that run off of either Excel spreadsheets, or glomped together masses of Salesforce customizations. I'm not convinced that this is an improvement in matters. I also think that there's just a lot of custom software that doesn't get built.
- bzzzt 6y ago
  I believe the web has killed a lot of those 'single user' Excel applications. If they had value, they would probably have been reimplemented in the back-end of new web-based business services (if you're unlucky by 'enterprise programmers') around the turn of the century.
anthonygd 6y ago
> As a result, the industry shifted to encouraging the use of "real" programming languages for everything, even the simplest systems, because the greater up-front costs were saved many times over by having a sane story for long-term maintenance and improvement. That makes a certain amount of sense; I know I've sworn at insane Excel spreadsheets that have lived far longer than they should have. The thing I'd argue is that the vast majority of business tools are only useful for very few people for a very short period of time. Sometimes they are never useful. Doing those in a "real" language can have an ROI of centuries (or never), so just never get automated or tried. There's a very important place for cheap prototyping and if non-developers can do it that's even better.

thrower123 6y ago

Is this really the case, or does stuff just continue flowing into the tools that are still accessible to users? I have seen an incredible number of businesses that run off of either Excel spreadsheets, or glomped together masses of Salesforce customizations. I'm not convinced that this is an improvement in matters. I also think that there's just a lot of custom software that doesn't get built.

bzzzt 6y ago
I believe the web has killed a lot of those 'single user' Excel applications. If they had value, they would probably have been reimplemented in the back-end of new web-based business services (if you're unlucky by 'enterprise programmers') around the turn of the century.

anthonygd 6y ago

> As a result, the industry shifted to encouraging the use of "real" programming languages for everything, even the simplest systems, because the greater up-front costs were saved many times over by having a sane story for long-term maintenance and improvement. That makes a certain amount of sense; I know I've sworn at insane Excel spreadsheets that have lived far longer than they should have. The thing I'd argue is that the vast majority of business tools are only useful for very few people for a very short period of time. Sometimes they are never useful. Doing those in a "real" language can have an ROI of centuries (or never), so just never get automated or tried. There's a very important place for cheap prototyping and if non-developers can do it that's even better.