I don't really get the hate. Everyone here uses AI for just about everything yet they almost never credit the models. I think this misrepresents the work so I credited the tools. But I'm by no means just handing a blank script to a model and hoping for the best.
> TiddlyInstall is a single HTML file. This page carries the installer programs, the runtime catalogue, and all the code that makes, edits and signs them
Author here: TiddlyInstall is based on the observation that programs written for runtimes like Python, Nodejs, and so on -- all follow the same basic requirements for installation. In that they need the runtime installed and then the software. Existing approaches like freezing apps and trying to port it to an exe directly usually break.
The simplest possible approach is to install the runtime on a users computer, isolated, with the package inside it. It's wasteful in terms of disk space but can be generalised to build reusable installers that work for just about any language.
Reusable installers
A reusable installer reads its own meta-data to decide on the runtime to install and the project. Linux and Mac are supported through a bash script while all windows versions use an NSIS program. The unusual thing about this is once you build the reusable installers since it's just meta-data everything can be edited inside the browser without compiling new binaries.
Like TiddlyWiki: You can actually download this entire system as a single HTML page to create installers without ever needing a server. Though Github project support only works online at the moment (not a limitation, haven't added this feature yet.)
> Designed by Matthew Roberts and implemented by Claude. If you like this software why not hire me?
Why would I hire you when I could pay $17 for a Claude subscription?
By "design" I mean choosing how the software works and the engineering trade-offs. I didn't type "make install system" into claude. I designed how the software worked based on my past research and my engineering knowledge. This included the original NSIS code I wrote: https://github.com/robertsdotpm/win-auto-py3 and file-based meta-data data here: https://robertsdotpm.github.io/software_engineering/program_...
I don't really get the hate. Everyone here uses AI for just about everything yet they almost never credit the models. I think this misrepresents the work so I credited the tools. But I'm by no means just handing a blank script to a model and hoping for the best.
> One file, like TiddlyWiki
> TiddlyInstall is a single HTML file. This page carries the installer programs, the runtime catalogue, and all the code that makes, edits and signs them
Click save page, buddy.
Author here: TiddlyInstall is based on the observation that programs written for runtimes like Python, Nodejs, and so on -- all follow the same basic requirements for installation. In that they need the runtime installed and then the software. Existing approaches like freezing apps and trying to port it to an exe directly usually break.
The simplest possible approach is to install the runtime on a users computer, isolated, with the package inside it. It's wasteful in terms of disk space but can be generalised to build reusable installers that work for just about any language.
Reusable installers
A reusable installer reads its own meta-data to decide on the runtime to install and the project. Linux and Mac are supported through a bash script while all windows versions use an NSIS program. The unusual thing about this is once you build the reusable installers since it's just meta-data everything can be edited inside the browser without compiling new binaries.
Like TiddlyWiki: You can actually download this entire system as a single HTML page to create installers without ever needing a server. Though Github project support only works online at the moment (not a limitation, haven't added this feature yet.)