In addition to Paxdiablo's computing centred list it might be worth to add that it's more of a linguistic issue than a technical or computing one.
The Slash was quite present way before data processing, back then known as Solidus or Oblique. Solidus, Latin for Shilling, as it grew out of the long-s, a letter, lost to modern Alphabets, used separate Shilling from Pounds. The long-s looks a bit like a stroke with a hook. The hook got lost over time due sloppy writing, transforming it into a straight slash. Solidus is in fact the name used within the Unicode classification for what we commonly call a slash.
Development and use case goes roughly:
Slashes have been used as number separator long before Comma and Dot settled like today. This dates back to at least the 16th century.
The usage is originated in the fact that in handwriting spaces can be rather random in width and appearance. Thus, separation is needed/helpful, some symbol gets inserted. Either a specific, or a generic like a slash.
In fact, in many early Latin texts and many medieval handwriting spaces were optional or omitted at all. If a word separator was needed, like to switch for numerals, a middle dot was common.
Slashes are still common in some countries (e.g. USA) as date separator (YY/MM/DD).
It was and is used wherever one wants to concatenate items but still mark em as distinct. Like the PL/I example, were it serves to enable reading 'I' as one instead of the letter I.
Similar the use with other product names like System 360 abbreviated to S/360. This continues with S/360-67 as abbreviation for the System 360 Model 67. Here a hyphen is used as secondary divider
Usually these applications are about saving space, so the slash may as well serve as replacement for various words. They can stand for and, or, for or many others.
This essentially works like the title case used in English language publications. All words not capitalized are candidates to be left out at whole, or replaced by a slash. Like Control Program for Microcomputers becomes CP/M.
In addition, much like there isn't one title case, but each publication fosters their own, people tend to make up their own variation - for example it's common to write simply 360/67 instead of S/360-67.
Speaking of, even companies change mid way, as there are as well IBM publications writing like that - or turning everything upside down by creating z/OS for something that started out as OS/360 :))
Of course all of this is in addition garnished with heaps marketing related stylization, great for flashy advertisement and even better to play trademark games.
The whole thing goes not only for products, but company names as well, or who remembers that M/A/I (sometimes M|A|I, today MAI) originally stand for Management Assistance Incorporated?
Long story short: Once it was about abbreviations, today everything goes, use it whenever pleases your intention.
Now, when looking at computing in particular, then there is a clear lineage, at least for some of the usage:
IBM stylized PL/I to fit the scheme started with the together with the S/360
Which in turn may have originated in use by US military bureaucracy like Another-Dave pointed out.
Intel's PL/M is a simplified implementation of PL/I, so it's name follows the precedence set.
PL/M was written for Intel by Garry Kildall. When creating CP/M, not long after, the naming choice was obvious.
Not much sleuth skills needed to see the continuation in MP/M and all that followed in DR's timeline of OSes.
S/360
seems to be a favorite pattern (general/specific) of IBM, they also hadS/34
and all their other names likePS/2
,OS/2
,PL/I
. But, of course, they seemed to brreak that withz/OS
for some reason.