VesperNet History,
section Preface


What about Macintosh Garden?

I took over the reins around 2014, lifted all limitations, poked the code here and there over the years to ensure that it would continue functioning well into the future. Due to design decisions of my predecessors, there are plenty of limitations and minor annoyances - though nothing that can't be worked around.

Over the years, the limitations have become more visible, and at some point I started a complete re-write of the site - I mean, how hard could it be? Right? Well, not hard but rather something that consumes a lot of time while maintaining the current instance of the website.

There is a lot of data, and there is a lot of data missing or saved in the wrong place, just because there isn't anywhere else to place said data. Apparently the amount of metadata for files and titles is severely lacking, and is a cause for annoyance.
This being something that my complete rewrite of the website would remedy.

In any case, an Alpha version was rolled out to test some of the new features and showcase a fraction of my vision for the website. Something that is available for and renders fine on both old/vintage machines as well as modern ones.
This Alpha version was written in PHP, and has since been severely outdated since I decided to switch it up and port it over to Python due to hitting a few walls and limitations with PHP.

Now, the backend is incredibly advanced, something that really doesn't show on the frontend. And at the time of writing, the final version of the document layout for software titles are being finalized, the sheer amount of metadata that can be available for a title can be either satisfying or overwhelming - leaning on the overwhelming side of things.
Though, even if it may seem overwhelming at first, the amount of data required to insert a new title is incredibly small, and can of course be further improved later by updating the fields whenever there is energy to do so - The biggest issue I'm facing is the UX, it needs to be presented in such a way that it satisfies the majority of users.

Since the current website contain a lot of data, I have written utilities to export both nodes and comments into JSON documents, and tools to load that data onto the new platform, but the final test and import would have to run after the finalized document structure. There is quite a lot to take into consideration, especially since we carry multiple pages for the same software but with different versions attached.

Another thing that requires a lot of attention are the files attached to each page, the current website puts them all in one folder depending on what it is, for instance all pages with the type app will put all application files into the apps folder, game type pages puts all files into the games folder, manuals into manuals for both apps and games, screenshots and thumbnails into one folder etc., you probably get that this isn't feasible in the long run.
So, it becomes a project in itself to move files around as well, even though this can be automated to some extent with the migration toolkit I've written.

The migration toolkit is something that deals not only with the metadata, but is also capable of bringing along all the files and somewhat sort them into the appropriate folders. This will of course ensure that we get rid of unwanted stray files and orphans, save us a few megs of space. Of course, the process must be rock-solid and can't be half-assed, so it is tweaked every now and then and is being continously tested to ensure that what we don't end up with missing data.

As such, I have decided to run the new instance and the latest snapshots live on VesperNet, so it would be a requirement during development to use the DNS to reach the dev snapshots and test things.
It won't be looking as "nice" as the current Alpha snapshot that is available on the public internet, but that doesn't matter at this time since, well, I need it to function as intended before applying polish.

I might have a vision, but I am no stranger to feedback; good or bad.