I was downloading some web files and discovered a feature in Windows that I had never noticed before. Windows automatically links a .html file to a directory with matching name and _files suffix. This is to allow you to save webpages and have the html file itself linked to the resources. This feature is called Connected Files and was first introduced in Windows 2000.

This works when the file extension of the main file has an .html or .htm extension. I tried some other web-related file extensions and they do not work. The attached resource directory can be named ending _files, .files, or a localised suffix like _fichiers. The intention was to allow browsers like Internet Explorer to save webpages for offline reading while retaining the ability to move files around directories. It still works when saving complete webpages using a modern browser like Vivaldi, because the feature is built into the windows shell. And so it also works if you manually create new files with the correct name patterns.

Delete, Move, Copy, Cut & Paste operations done to one operate on the other as well. Rename operations cause a warning to appear that the linkage will be lost. Only operations run within a normal shell will allow linked operations to happen. I.e. it needs to happen through windows explorer, the use of a command shell (cmd or powershell) or other application won’t allow linked operations.

It can be turned on and off using the following DWORD registry key:

HKEY_CURRENT_USER\SOFTWARE\Microsoft\Windows\CurrentVersion\Explorer\NoFileFolderConnection

Set to 1 for disabled and 0 for enabled.

I got some of this information from a very old MSDN Magazine article published in November 2000. Thanks microsoft for keeping it archived and still accessible: https://learn.microsoft.com/en-us/archive/msdn-magazine/2000/november/windows-2000-registry-latest-features-and-apis-provide-the-power-to-customize-and-extend-your-apps

I think it’s neat that I am still discovering stuff that was built decades ago, and it’s great that it is still works and is actually useful. Some other apps have similar capabilities that can interpret these files linked by name patterns, for example Calibre will accept an imported html file, and it will copy in the linked resources.