OS not web browser
I’d say it is not the web browser because the APIs used by most programs generally end up using the APIs provided by the OS.
It is easy enough to check because the web browser will be using a Web API (e.g. Firefox Web API for file selection) that is different to the API used in other applications (e.g. Microsoft Word). A programmer with the inclination could create a Javascript routine to test file selection in both applications which run Javascript.
I just did it by selecting images to upload in Firefox and to insert in Microsoft Word - all in WIndows 10. The file selection behaviour is the same in both.
-
Discourse in Firefox
-
Microsoft Word
File selection behaviour is unusual but not the complete problem
The files are selected in such a way that the last file selected always appears first in the file list so you only have to look at the start of the list to check which file is currently being selected.
- “file 1”
- “file 2” “file 1”
- “file 3” “file 1” “file 2”
- “file 4” “file 1” “file 2” “file 3”
As long as the files are selected in the same order then the list of files looks the same whether I use shift to select a list or ctrl to select each file individually or a combination of both.
It could have been done differently but I like seeing the last file selected at the start of the list because it means less eye movement for me. However this has two problems:
- It is counter-intuitive not to see the last file selected at the end of the list.
- That list sequence is probably not what is used to provide the sort order in Discourse despite what we’ve been told.
I’m happy to be proved wrong.
Display order in Discourse appears to be based on something like upload completion times
The file order appears to be based on the sequence of completed uploads. It didn’t matter how many ways I selected the files, the order on Discourse was usually different. This is a common problem on many other sites and networks (e.g. print spooling) where upload completion sequence determine the storage and display order.
Here’s examples of what I got on Discourse for image files originally created in a time sequence from 1 to 4.
Selected individually in reverse order
Run 1
Run 2
Selected as a group in reverse order
Run 1
Run 2
P.S.
The file size sequence is 1, 4, 2, 3 from smallest to largest. Interestingly enough that is the reverse of what appears to be the most likely base sequence above 3, 2, 4, 1.
The file data by date and time is as follows:
20/08/2018 08:50 PM 550 1.png
20/08/2018 08:51 PM 697 2.png
20/08/2018 08:51 PM 809 3.png
20/08/2018 08:51 PM 600 4.png
P.P.S. Yes, I know that I shouldn’t have used the reverse sequence - it just complicates this. I only did that because someone suggested that would be a solution to the problem.
