3 ms·

There will be code

10 points by nreece 18y ago

Zak 18y ago
I had to look up MDA just to see what it is. It sounds like one of the technologies people making "enterprise" software would use. It does not sound like it would actually make software easier to create. That said, I do think that visual programming languages have merit, but they require the same sort of logical thought process as any other sort of programming. What the article seems to be arguing against is that some magical technology will come along and make it so that programmers aren't needed. I don't think anything short of human-level AI will do that. COBOL was supposed to do that.
ojbyrne 18y ago
The products we build are quite different. Someone invented GUIs to take up developer's time. Instead of simple procedural programs that start at the top and run to the bottom (with some loops and conditions) we have event driven, multi-window, complex applications.
gengstrand 18y ago
Remember all those graphical workflow generators such as http://www.activevos.com/community-open-source.php http://www.activevos.com/community-open-source.php and http://www.oracle.com/technology/products/integration/workflow/workflow_fov.html http://www.oracle.com/technology/products/integration/workfl... that were supposed to replace the need for coders? I don't have a problem with EAI. I think it's great. I just don't agree with the overselling of what EAI can do.
JunkDNA 18y ago
Early stage computer science research is not an area I'm familiar with, but I would be curious to know where the current research is headed in terms of radically rethinking the way we think about and represent code. That said as an admittedly very naive observer, it would seem to me that it's very difficult to get away from the basic building blocks we've come to know, not only from a social standpoint (i.e. several generations of programmers think in these terms now). But also because the hardware hasn't changed fundamentally. While capacity and compute speed have gone up, today's computers still work the same way at their core as their predecessors. The math they use is the same, the basic components are the same (until memristors hit the scene I guess). The operations microprocessors can carry out and the way they need to access data forces you to implement the core of a programming language a certain way. Once that happens, then all the abstractions on top of that start to inherit many of those semantics. When you get right down to it, processors shove binary data around and permute it based on a set of well-defined mathematical principles. Until they start doing things differently, I feel like we're "stuck" with those "if" statements (or their logical equivalents) and all their cousins for a while.

10 points by nreece 18y ago

Zak 18y ago

I had to look up MDA just to see what it is. It sounds like one of the technologies people making "enterprise" software would use. It does not sound like it would actually make software easier to create. That said, I do think that visual programming languages have merit, but they require the same sort of logical thought process as any other sort of programming. What the article seems to be arguing against is that some magical technology will come along and make it so that programmers aren't needed. I don't think anything short of human-level AI will do that. COBOL was supposed to do that.

ojbyrne 18y ago

The products we build are quite different. Someone invented GUIs to take up developer's time. Instead of simple procedural programs that start at the top and run to the bottom (with some loops and conditions) we have event driven, multi-window, complex applications.

gengstrand 18y ago

Remember all those graphical workflow generators such as http://www.activevos.com/community-open-source.php http://www.activevos.com/community-open-source.php and http://www.oracle.com/technology/products/integration/workflow/workflow_fov.html http://www.oracle.com/technology/products/integration/workfl... that were supposed to replace the need for coders? I don't have a problem with EAI. I think it's great. I just don't agree with the overselling of what EAI can do.

JunkDNA 18y ago

Early stage computer science research is not an area I'm familiar with, but I would be curious to know where the current research is headed in terms of radically rethinking the way we think about and represent code. That said as an admittedly very naive observer, it would seem to me that it's very difficult to get away from the basic building blocks we've come to know, not only from a social standpoint (i.e. several generations of programmers think in these terms now). But also because the hardware hasn't changed fundamentally. While capacity and compute speed have gone up, today's computers still work the same way at their core as their predecessors. The math they use is the same, the basic components are the same (until memristors hit the scene I guess). The operations microprocessors can carry out and the way they need to access data forces you to implement the core of a programming language a certain way. Once that happens, then all the abstractions on top of that start to inherit many of those semantics. When you get right down to it, processors shove binary data around and permute it based on a set of well-defined mathematical principles. Until they start doing things differently, I feel like we're "stuck" with those "if" statements (or their logical equivalents) and all their cousins for a while.