Shopping-mall closed as warning message on glass door in Boston:
https://st.depositphotos.com/1005125/2506/i/950/depositphotos_25060035-stock-photo-shopping-mall-closed-as-warning.jpg
Shopping-mall closed as warning message on glass door in Boston:
https://st.depositphotos.com/1005125/2506/i/950/depositphotos_25060035-stock-photo-shopping-mall-closed-as-warning.jpg
Not directly relevant but I remember being given the source of a callable routine written in Assembler and the number of errors was approximately equal to the number of lines of code, in the region of 10 - 20.
This was 40 years ago and I can't remember all of the details.
Not directly relevant [...] Assembler and the number of errors was approximately equal to the number of lines of code, in the region of 10 - 20.
I think I have that beat. Back in the late '70s as a student for one course I had to use punched cards to submit the programmes, however I think we were the last year to need them for that programming course - they went on to ASR33 teletypes. Consequently, the maintenance was rather spotty, and in particular the ink ribbons were almost always nearly dry and hardly made a mark, and there was often a queue for the odd machine with a better ribbon, though I'd happily use a different one and just peer closely. One program I entered and submitted for run was about 2-3 pages of well-commented code, and should have had an output less than a single sheet. I eyeballed the printed line along the top of the cards and they looked OK, but when I got the job back, it was an about 1" thick stack of paper filled with compiler errors, with quite a few 'too many errors on line, ignoring' to keep things somewhat in check. When I tracked down the problem, it turned out to be a single character wrong, and not a wrong key pressed, but a shift not applied - I wanted a / but got a V, these being on the same key (upper case only!, / was shift-V). This meant that the close comment */ was actually *V (the RHS of the V was much more visible than the left which didn't mark at all, so it looked like the / I wanted), and since this comment was at the end of the typedefs, it missed the closing of the typedefs and starting of the next section, the compiler took the variable declarations as malformed typedefs, the actual code as malformed variable declarations, and never got round to wondering where the code was as it had hit the limit on listing pages of output. I repunched that one card and everything was fine, and IIRC there was only a single hole different. This was 45 years ago, and you can tell it made an impression on me.
This reminds me (peripherally) of an old saw amongst programmers, or at least ones that I know:
Every non-trivial program contains at least one bug and at least one extraneous operation. Therefore, every program can be reduced to a single line of code, which won't work.
Not directly relevant but I remember being given the source of a callable routine written in Assembler and the number of errors was approximately equal to the number of lines of code, in the region of 10 - 20.
This was 40 years ago and I can't remember all of the details.
We had one similar when I was taking a programming course.
We were using old school COBOL, and keypunch cards that were fed into an IBM360 mainframe. And the gal that was in charge of the typing class managed to get access to the keypunch machines at the college for extra credit for her students.
We would hand in our 80 column worksheets, and a few days later get them back with the cards. But the problem was, they were using common typing shortcuts of the time. Like using lower case "l" instead of the number "1", and the upper case "O" instead of the number "0". So basically every one of the cards they had typed was useless.
I saw, on a cash register, decades ago:
'Sory, we not cashe cheks.'
That about took the cake.
On the other hand, I have far more sympathy for hand-lettered (or quickly typed) signs than I do for things that have been professionally printed and should have had multiple people read them along the way. Those annoy the heck out of me.
One of the more recent: an elevator I rode a while ago had the following on an engraved metal sign:
'Certificate of Complianace is located in Management Office.'
Googling it, the sadder part is that 'Complianace' gets plenty of hits (once you tell Google that, yes, you want that spelling) from people who really should know better.
Certificate of Complianace
That's a weird one since it clearly differs from the spoken word. Even the worst writer should notice the wrong spelling. It was probably a non-native English speaker just learning the language who wrote the note.
That's a weird one since it clearly differs from the spoken word.
Perhaps the wording was sent to the engravers by mobile phone.
AJ
That's how I felt, too - until I went through the Google hits. Plenty of US people (and businesses!) spelling it that way.
Here's an example: http://duboispa.gov/wp-content/uploads/2013/10/Application-for-Smoke-test.pdf
Both the incorrect and correct spellings on what appears to be a city government form.
We need to be a bit careful around this topic though, I've seen stories here - and story descriptions, blog entries - which look as though they were translated by Google Translate, anno 2000 at the latest.
which look as though they were translated by Google Translate
Too bad so many people use google translate. It's definitely not the best translator out there. You have to look for a language specific online translator and those are often way better than anything google produces.
You have to look for a language specific online translator and those are often way better than anything google produces.
That is difficult when you don't know what language it is. Google Translate has a detect language feature for the source.
That is difficult when you don't know what language it is.
It's trivial to find out what language it is. Type 'translate' and a few of the words to translate in any search engine and you are very likely to see what language it is just by reading the descriptions of the first results.
Google Translate has a detect language feature
It's an automatic feature on Twitter. A response of 'OK' to one of my posts (according to Google) was the Arabic for 'OK' :-)
AJ