The following is one of a series of posts involving the use of AI for VFX and other development. This will be spun off to its own blog soon, I hope.
The reason I mention these things (the problem of the development environment and framework choice) is based on the idea that I can contribute real experience and insight into new technologies for the benefit of myself and my friends. If that is not true, if I explain myself badly or if people just do not believe me, then I am wasting my time. In this case I think I was frustrated by another example of the situation I describe below. It is not a waste of my time because even if my friends think I am a moron, I still have to learn these things.
The short version is that in software development, there are many, many choices to make before you can write much code (or use or test something Claude does). Since the internet explosion of roughly early this century, that has become an adventure game with lots of twisty passages all alike.
However, even here Claude may be a game changer.
The longer version goes like this. It has always been the case that when starting a software development project, that one must put a development environment in place. Not all elements of that software development plays nicely with all other parts, that would be way too expensive and is usually not necessary. It is also the case that one has to constrain a problem to fit the resources at hand and failure to make those choices wisely can wildly screw up a project. Generally many if not most of these obstacles can be finessed with small amounts of money, it is not a big deal. Sometimes it is very expensive, but that is a fact of life. (Example you want to develop a lisp application for Symbolics but you dont have access to a Symbolics machine).
Since the early 90s this development environment issue has been one of the high order bits of my ability to be productive. Some of these problems are trivial (e.g. you want to develop for iPhone, so you need a mac and an iPhone). Some of these are annoying, you have to use a specific development environment and you discover that it requires a later version of the OS and a several week learning curve (or worse). Therefore, a reasonable development plan, even with Claude, is to either use a development environment you know well, or budget in the time to figure out what that development environment needs to be and see what budget, etc, is required.
In this case, my unofficial advisor, Henry Fuchs, came up with an entertaining, non trivial app for a phone he would like to see me write with AI. He has an iPhone, I have an Android. Claude can handle this very easily, but obviously there are framework implications. I need a separate path for the iPhone from the Android, and a plausible framework for cross platform development. The Claude chosen framework for this is called Flutter from Google, it is free, it should work on all platforms, and there is some configuration problem on my Windows 11 box and/or the current Windows Subsystem for Linux (which is excellent by the way) that is causing it to fail.
I am out of practice, I forgot about these issues in a desire to impress Henry. I am bad, no doubt, but the problem is a cost of doing development.
The good news is that I should be able to give Claude direction as to which development environments and frameworks I want to use.
