Securing Electricity Meter Cupboard

[stillp]

How does the OED define "verbose?"

And why doesn't it define 'gullible'?

 

[stillp]

"As an even more egregious digression, am I the only one who sometimes looks at the list of suggestions offered by a spelling checker, and wondered WOE the person who wrote it did not consider the possibility that the writer may have simply hit a key vertically, horizontally or diagonally adjacent to one which he should have?"

I believe the alternatives offered are those that users who have signed up to the "help Microsoft improve..." program eventually settled for, i.e. the software collects the user responses and uses those to amend the internal dictionary.

 

[JohnW2]

I believe the alternatives offered are those that users who have signed up to the "help Microsoft improve..." program eventually settled for, i.e. the software collects the user responses and uses those to amend the internal dictionary.

Maybe, but if not the original author of the software, I suspect that those users must have been thinking about key positions on a keyboard. As I just wrote, I seriously doubt that (m)any people who didn't know about keyboards would even suspect that "ijcoereft" was meant to be "incorrect".

Kind Regards, John

 

[stillp]

No, but if a number of people had corrected "ijcoereft" to "incorrect" then the software would have learnt that.

 

[JohnW2]

No, but if a number of people had corrected "ijcoereft" to "incorrect" then the software would have learnt that.

Agreed, but the probability (sorry EFLI!) of anyone making that particular combination of typos when trying to type "incorrect" must be incredibly small (and the number of possible mis-typings of "incorrect" with 3 or 4 'key errors' extremely large). Even with my pretty awful typing, I don't usually hit the wrong key more than once or twice in any one word

There is also, of course, the possibility that a number of other people will have corrected "ijcoereft" to something other than "incorrect" - which would rather muddy the waters.

In any event, BAS seemed to be implying that he didn't think that spell-checkers were very good with 'nearby key errors'. I think I've at least partially disproved that belief, regardless of whether it is coping with such errors because of the way the algorithms were originally written or because of subsequent 'learning' on the basis of user input.

Kind Regards, John

 

[JohnW2]

No, but if a number of people had corrected "ijcoereft" to "incorrect" then the software would have learnt that.

Another interesting thing (I would say an 'undesirable' functionality) about the Word spell-checker is that is seems unable to deal with any mis-spelling of a word that has been user-added to its dictionary....

... I have just added "presbyopia" (which wasn't previously recognised) to my Word dictionary. It now correctly recognises that word if I type it correctly, but does not suggest the correct spelling (or any other spelling) even if I make a single simple typing error - even a phonetic one like "prezbyopia".

Kind Regards, John

 

[ban-all-sheds]

In any event, BAS seemed to be implying that he didn't think that spell-checkers were very good with 'nearby key errors'.

am I the only one who sometimes looks at the list of suggestions offered by a spelling checker....

How you can infer that I was, or might have been, saying that they all have that problem, I really don't know.

I think I've at least partially disproved that belief, regardless of whether it is coping with such errors because of the way the algorithms were originally written or because of subsequent 'learning' on the basis of user input.

Next time I encounter an example, I'll hopefully remember this, and come and tell you about it.

 

[JohnW2]

... I have just added "presbyopia" (which wasn't previously recognised) to my Word dictionary. It now correctly recognises that word if I type it correctly, but does not suggest the correct spelling (or any other spelling) even if I make a single simple typing error - even a phonetic one like "prezbyopia".

I've just realised that my previous comments all related to the Word 2000 spell-checker. The Word 2010 one is a bit different, in at least some senses 'inferior' ....

1...In the case of "incorrect", it cannot cope with as many 'nearby key' errors as can Word 2000. It does not make any suggestions for "ijcoereft".

2...The Word 2010 dictionary already contains "presbyopia". However, again, it can only cope with a couple of errors - "peezbyopia" is an example of about the worst for which it suggests "presbyopia" (or anything).

3..."hypermetropia" was not in the Word 2010 dictionary, so I added it. Unlike the Word 2000 checker, it was then capable of detecting errors in that word, but, again, only up to a maximum of a couple of errors - "hypernetrooia" is an example of about the worst for which it suggests "hypermetropia" (or anything). However, it can also cope with at least two missing characters as well as two incorrect ones - it does suggest "hypermetropia" for "hypernetroi"

However, I've also discovered that what the Word 2000 checker did with "incorrect" is probably absolutely nothing to do with it thinking about "nearby keys". Whilst it is true that it suggested "incorrect" when I typed "ijcoereft", it also suggested "incorrect" when I substituted fairly random characters (not 'nearby keys') for all the erroneous ones - e.g. "izcoqremt". Hence, whilst is seems pretty good at detecting multiple incorrect letters, despite what I initially thought, that doesn't appear to be anything to do with it 'thinking about nearby keys'.

There is obviously going to be a lot of difference between different spell-checkers. For example, the one used in this forum suggests (only) "therefore" for "ijcoereft" and (only) "recommit" for "izcoqremt", although the Word 2000 dictionary suggests "incorrect" (which is what I 'intended') in both cases.

Kind Regards, John

 

[JohnW2]

Next time I encounter an example, I'll hopefully remember this, and come and tell you about it.

Thanks - but, as I've just written, it appears that the spell-checker I was examining does not 'think' about 'nearby keys' - once a character is incorrect, it seems to make the same suggestion(s) whether the erroneous character relates to a nearby key on the keyboard or a randomly-chosen one whose key is at the other end of the keyboard. However, that's just one checker - there may be some which do 'think about nearby keys'.

Kind Regards, John

 

[ban-all-sheds]

Maybe/probably.

The thing is, you wouldn't know, for if you make a miss-hit error and the checker offers you the right word, it would have done it's job, and you'd not know how it had done it. It's only when it doesn't offer you the right one that you realise it's trying to find one close to what you typed without considering the possibility of an adjacent key miss-hit.

 

[JohnW2]

Maybe/probably. ... The thing is, you wouldn't know, for if you make a miss-hit error and the checker offers you the right word, it would have done it's job, and you'd not know how it had done it. It's only when it doesn't offer you the right one that you realise it's trying to find one close to what you typed without considering the possibility of an adjacent key miss-hit.

Well, I've now played a fair bit with the two Word spell-checkers, and have yet to find any case in which it suggests the correct word when there is an 'adjacent key error' but does not suggest the same correct word when one changes that erroneous character to anything else incorrect (not 'adjacent key').

That strongly suggests to me that it is not considering 'key proximity' - or that, if it does, considering it does not make it any more likely to make the right suggestion than would be the case if it didn't consider key proximity. Do you agree?

Kind Regards, John

 

[ban-all-sheds]

I think it's a sound conclusion.

I've not done any testing, but I have, frequently enough for it to register as something which often happens, noticed it not suggesting the word I wanted when that word is both incredibly commonplace and is found by considering adjacent key miss-hits.

 

[ban-all-sheds]

... incredibly commonplace ...

Now there's a thing.

When did a word meaning not believable come to mean very? I could have just as well written "unbelievably commonplace" and prompted the same question.

 

[JohnW2]

I think it's a sound conclusion. I've not done any testing, but I have, frequently enough for it to register as something which often happens, noticed it not suggesting the word I wanted when that word is both incredibly commonplace and is found by considering adjacent key miss-hits.

That's certainly consistent with my conclusion in relation to the two spell-checkers I've looked at.

Unless you had errors in multiple characters, I think you might find it hard to draw any conclusions. With the checkers I'm looking at, in the great majority of cases, any single-character error will result in the correct word being suggested - for example, I can replace any character in "incorrect" with a "z" and it will always suggest (only) "incorrect". Even if the checkers include functionality which enables them to consider 'adjacent key errors', they therefore seemingly have the ability to make exactly the same (correct) suggestion without needing to utilise any such functionality which may exist.

Kind Regards, John

 

[JohnW2]

Now there's a thing. When did a word meaning not believable come to mean very? I could have just as well written "unbelievably commonplace" and prompted the same question.

Quite so - it's what is known in the trade as "evolution of language"

Of course, there are a whole pile of similar adjectives/adverbs for which much common usage has evolved well away from what the words actually mean/meant ... terrible, horrible, amazing, fantastic etc. etc. etc. (and all the corresponding adverbs).

Kind Regards, John