Cisco IOS - its the reason I never went very deeply into configuring network equipment. The few times I had contact with IOS resulted in much hair pulling and the sort of language you don't want your boss to hear.

if he were Mycroft then this particular episode would never have occurred.

I didn't notice whether Watson, not being the brains of the firm, found the solution or someone read the manual.

net no sh ip xxx -- No shit Sherlock.

Network engineers never were my favourite IT people partly because of this sort of brain damage. "No shut = open" verges on Newspeak double plus ungood.

Just wait till you hear about "switchport trunk allow vlan" sans add.

Good times.

Well that's my submission into "Who, me?" scuppered!

> Network engineers never were my favourite IT people partly because of this sort of brain damage.

That's why good ones earn a packet....

as a former "cisco network manager" i can relate to this; been there, seen it, got the tee-shirt !!!

ios is great - until you screw up. one command is all it takes to bring an entire global network down (don't ask how i know that one !!)

”sh” will resolve to ”show” everywhere where there isn’t a better match, such as in interface configuration mode, where ”shutdown” is available. So in this case, Sherlock would have simply run the ”shutdown” command on a remote interface, And since in classic IOS commands were executed instantly - branch, meet saw.

One should note here that the command to do something on a Cisco is never ”sh”, there’s always a full, expanded command (”show”, ”shutdown”, or even ”shim” in some places) being executed (even if you don’t see it at the prompt). It’s the console cowboy’s responsibility to know what he’s typing. And believe me, I’ve mistyped a lot of IOS/XE/XR commands over years.

Juniper’s JunOS CLI, on the other hand, auto-expands the commands as you type so you’ll know what the router thinks you meant to do before you hit enter. (You still need to know what you’re doing, though!)

Oh yeah. It was way back when Lotus Notes was Release 4. LotusScript had just appeared and I had been tasked for creating a Content Management System for an international customer.

Now, you need to understand two things : the first is that, in those days, a Notes database was limited to 2GB in size. The second thing is a little quirk in the system when you're cycling through all the documents (or records, for you RDBMS people) in a particular view.

The manual says (still to this day) that using view.getnextdocument(olddocument) will give you the next document in the view, which it pretty much does. What the manual does not say is that the old document is not deleted from memory when you do that.

So, in those days, it wasn't much of an issue because Notes databases did not contain hundreds of thousands of documents.

But then R5 came out, and the 2GB cap was lifted. And oh boy, did those databases start filling up. That is when I learned, the hard way, that cycling through 85K+ documents, with attachments, in a view could crash the script (thankfully, not the server). The bug report came in from the customer, and I spent days trying to understand why because the script never crashed on the same document (thank God computers are supposed to work on zeros and ones - God only knows what would happen if fuzzy logic were to be used).

In any case, after much head-scratching, I finally clicked that the script always failed after more than 80K documents had been processed. It took me another few minutes and then I wrote this :

Set olddoc = doc

Set doc = view.getNextDocument(olddoc)

Delete olddoc

With that, magically the script always completed successfully from then on (Delete removes from RAM, not from database).

That is one lesson I have never forgotten.

Backward double-negative logic is a predictor of massive future industry dinosaurs.

Fred Brooks called IBM's OS/360 a "multi-million dollar mistake" back when that was a lot of money. Now called Z/OS, its backward COND statement ("Don't run this step if the previous step's return code was not N) is still the mainstay of the financial systems across the world.

“Typing sh-TAB can therefore be a shortcut to the command show, thus saving two whole keystrokes and valuable microseconds for greater productivity.”

Well, really only one keystroke…

I had the same thought, but then I remembered that it will also include the space after 'show' or 'shut', so it's still a net two keystrokes, after all.

Sherlock-adjacent icon -->

Ouch... This felt like reading my own story from 1996. Done exactly the same by trying to get the interface information using the same cursed "sh" shortcut.

This sort of crap trap is the reason I got into the habit of typing commands in full - much to the amusement of my colleagues.

Nothing to with networks, but operating a 1.5 m (diameter) infrared telescope at an Italian observatory as an undergrad, I was rather horrified that when I typed in the wrong coordinates, and the system asked me to verify the coordinates were right (Y/N), my response "N", resulted in a cheerful "Then I go!" followed by the humming of drive motors to swing the telescope into the requested position. This could result in the liquid nitrogen and, worse, liquid helium to be poured out of the cryostat of the infrared spectrograph we were testing, potentially causing major damage. Happily nothing bad happened, but I did learn from the Italians running the place that any key except Ctrl-D would be interpreted as "Yes". Brilliant!

Suffice to say I was not impressed by their UI design.

With something like that the UI should always be designed to 'failsafe'. For the given example it should have been that ANYTHING other than a single key combination (e.g. Ctrl-Y) would be taken as a 'no'. It should never have been set so that the 'default' result was the potentially dangerous condition.