Topic kinda says it all...can we get a misspell check like firefox / MS Word that puts a squiggly red line under a misspelled word and when you right-click on that word it offers you some recommendations on the correct spelling? Of course, this would have to be done in the text box before you post it to the channel / query window otherwise it would be useless. This would rock.

Just an idea

- samten

Scripted spell check by your's truely. (Must be on windows, w/ MS word 2003 or later installed): My SpellCheck

ah bummer. I don't have Word on this computer

Oh, Thought you did after mentioning MS Word in your post. I also meant 2003 or later...

It's cool. I'll just dig up my copy of it. I have it around here somewhere, I just have yet to put in on my new computer here. Can't wait to use your spell checker

I use this script as I don't have any MS Word products installed, rather nice script runs perfectly, no performance impacts.

Edit,
Works in 6.34 - 7.22 (version at time of this post)

Funny, I actually helped SwitchDragon write that addon I used to be a part of ProjectX, with SwitchDragon, Urmac, FolkLore, Mentor, UnConeD and all the others lol. Just brought me down memory lane, thanks

I've been looking for a spell checker for a long time now, still yet to have found a decent one.
All the current scripts are obtrusive or simply don't work any more (sometimes both).
A hunspell/iSpell integration for mIRC would be greatly beneficial for everyone (similar to Chrome's spell check - simply marks the misspelled words with a red underline and gives you the correct spelling when clicked).

The main problem with that... there isn't any underlining available on the edit line. That would have to change. Besides, IRC is just for chatting. There isn't any more need for good spelling on IRC as there is on IM, where most people use shorthand "IM-speak" instead of typing out words anyhow. On a browser, spell checking makes sense because you might be filling out a resume or doing something else important. That just isn't the case with IRC. Besides, auto spell checkers dumb people down. You don't learn to spell if everything is done for you. You learn to spell by actually having to do so.

IMO the argument that we are making society stupider by introducing spell check is nothing more than anecdotal evidence and conjecture. I don't think it's mIRC's job to make opinionated calls on how people should or shouldn't use their computers. Telling people they get dumber by using spell check is not really helpful.

The real issue, in my opinion, is that to implement spellchecking in mIRC, Khaled would likely have to put in a disproportionate amount of effort for the return. The amount of false positives you get from a spell checker over a medium like IRC is likely ridiculously high. False positives are a big problem because they make users ignore feedback-- ie., your red underline would be meaningless if it popped up on every 2nd word. That's likely what would happen over IRC. How do you spell check a sentence like "u r gr8"? Those aren't "misspellings". How do you deal with the fact that many of the spelling errors would just be channel or nicknames, or shorthand notation for the various things people talk about over IRC?

I should also point out that I've used free spelling libraries like hunspell (an aspell-like library) in the past, and they are not very impressive. They tend to lack concepts of pluralization and verb tenses, so if mIRC were to add standard IRC jargon like "lol", "brb", etc., it would not handle all of the other possible permutations of these words: lols, loling, lol'd, etc... it's also important to realize that many spelling errors do not produce incorrect words, but rather, another valid word (known as a typo). Misspelling pig as pic, for instance, would not generate an error line, so this would be a false negative. With the possibility for so much misjudgment on the part of the spell checker, I'm pretty skeptical about adding this functionality.

As a final point, I think this image is useful to share, given the subject matter. The following is what my browser, Chrome, thinks of my spelling abilities. Chrome uses aspell, I think, which is what has been suggested for mIRC. Although I made no actual spelling errors in my post, Chrome seems to think there are plenty. FYI it doesn't even recognize "spellcheck" as a valid word:



That's that we'd be expecting to see, if mIRC implemented the same functionality. Except it would be worse-- I'm not nearly as formal about my spelling/grammar on IRC. I suspect other people have similar habits.

We already know that spell checkers are not perfect. You don't need to explain the downsides of using a spell checker. It doesn't change the fact that it is a popularly requested feature for mIRC and one that many people including myself would find useful in spite of those downsides. Spell checkers may not be compatible with the way you personally use IRC, but that doesn't mean it wouldn't be useful for others.

On a side note, as your own image shows, Chrome doesn't underline any part of "u r gr8", which sort of ruins the point you were trying to make there. In any case, if you are the type of person who uses terms like "gr8" or "ur", then you probably aren't the type of person who would enable the spell checker anyway.

Are you really speaking on behalf of every single reader, here? Most people aren't aware of the actual downsides of using a spellchecker in shorthand text mediums-- you know, ones where full sentences are not the norm (aka. IRC). I'm bringing this to light.



It doesn't really ruin it, no. You're forgetting an important point about mIRC. It's not Chrome. The dictionary mIRC uses may or may not catch those words. It's possible that Chrome specifically curates its dictionary to ignore words with numbers in them (something mIRC would have to mimick, unless the library does this automatically), as well as "words" that are 1 letter long. Note that I asked a question, I didn't provide an answer. The question was: "how will mIRC react to these sentences?" -- the answer is, it's hard to predict whether it will do the right thing, given an arbitrary example. Congratulations, Chrome (not mIRC) did the right thing for one of them. In English (would it work in Arabic? Russian?). That's a very specific and small victory.

Also, "if you are the type of person who uses terms like 'ur' ... you will disable it" isn't really the point. Again, "ur" is just one example of a false positive. You are assuming that all false positives come from someone who doesn't "care" about spelling to begin with. I gave many examples of when that would not be true. For instance, I care about spelling. I'm very conscientious about my spelling. And yet, in this reply alone, Chrome has made me question whether I spelled "mimick" correctly (I did, Chrome doesn't have the variant spelling in its dictionary), thereby wasting my time. This is fine and dandy on a forum where my answer will stick around, but on IRC my answer will disappear into the backbuffer in a few minutes time, so I would not waste my time googling to make sure mIRC is making the right call every time something pops up red.

Again, this is just me-- but the question is, how many people are like me? And how many people would actually benefit from a very imperfect spell checker? This seems more like a cost benefit issue to me. Yes, in a perfect world, we could add every single feature you wanted. This is reality, where it really doesn't make sense to spend a large amount of effort (the editbox would need to be completely redesigned, a spellcheck would need to be integrated, dictionaries curated in multiple languages) so that a handful of people catch 30% more spelling errors. If you can prove that the feature would be more effective/useful, that would be more convincing. Implementing something that is known to not be very effective seems... ineffective.

I disagree. Spell checkers are very prevalent in IM clients, so many people are already familiar with how they would work in such a situation.



Chrome's spell checker was correct. "Mimick" is an archaic spelling, so it is no longer accepted as a correct spelling in modern English. Similarly, from your last post, "spellcheck" as a compound word is not a widely accepted spelling either. But even ignoring these facts, nothing is stopping you from loading in a custom dictionary or manually adding those words to the internal dictionary so that they will be treated as correct.



That is something for Khaled to decide, IMO.

Everything is for Khaled to decide. That doesn't mean no one should comment on anything. I'm certain Khaled takes into account not just the suggestions, but the general opinions of those who post as well as any good or bad points that are stated regarding a specific feature. There really isn't anything wrong with agreeing or disagreeing with a suggestion and giving reasons why.

Regardless, this suggestion has been suggested so often that I think every pro and con has already been stated many times already.

The "short" answer for the OP - the suggestion is known and has been known to Khaled for a LONG time. It would require redoing how the edit line works to allow for underlining and hotlinking. Khaled would also need to either select a set of dictionaries that he thinks will never go away or else he will have to manage his own set. It doesn't help to use third party items in a program only to have the people who manage them disappear. Most programmers generally avoid that kind of situation whenever possible. Exceptions are of course for major companies (such as using .NET because you know M$ isn't going away). Maybe Khaled can find a dictionary set for most languages that he feels will be reliable and long lasting so he doesn't have to worry about it. That or he has to cater to the ability to let users pick their own dictionaries and that can be a pain as well. And considering scripting works (it may not be perfect, but it does work), I have a feeling the feature is far down on his list of things to do. That being said, I do believe he has it on his list of things to do. Of course, that's just a guess.

I'm glad you learned how to google and click the first result. Unfortunately, you're repeating an uncorroborated statement that has little bearing on reality. I can't find anything that backs up your claim that mimick is "no longer accepted as a correct spelling". When was that decision made? Was there a meeting? Mimick is still used in prose today, and the alternate form "mimic" is actually just as "archaic"; it dates back to the 1500's according to Google books search results, which, by the way, is actually prior to the spelling of "mimick" as these results show. So, contrary to your implication, "mimic" is not a newer version of the word mimick-- it's actually the other way around.



You're right about this, but this is where the complications I mentioned above come into play. Feel free to actually read what I wrote in my first post, but just in case, I'll reiterate: these libraries tend to be really bad at pluralization or verb tense/conjugation; they basically have no concept of it. You could add "foo", but "foos" "foo'd" "fooing" and other versions would still be considered incorrect. You'd have to add every permutation of every word outside your dictionary. You make it sound trivial-- it's not. I know because I've recently done this using hunspell (hence my mentioning of it above) for technical documentation I recently wrote. Technical documentation, with lots of technical jargon, fell outside the scope of the dictionary many times. Unfortunately, adding most of the technical words was not enough, as they were conjugated and used in many forms. We actually discovered that our custom word list ranged into the hundreds of words, and it took quite a while to sift through all the text to pick out those words (even though we had tools to help).

This is not a problem if the dictionary is good-- the problem is most dictionaries aren't... and they can't be, because of situations like technical terminology and even simple things like "mimic" vs "mimick" (neither is wrong, one is just no longer popular in the US, which, I should point out, is only one country of the other English speaking countries out there). But realize the point I'm making-- it's not about adding words. Because, rather than bothering to add these words, most people will be trained to ignore the red lines. This follows the same UI principles surrounding the overuse of confirmation dialogs; if you show a user too many of them, they begin to just click Yes without reading. False positives and inaccurate spell checking causes the same problem. The question is; how many false positives do you get over IRC; and how many spelling errors are being missed that would have been caught?

This goes back to the cost benefit question. Yes, it's something for Khaled to decide (I basically already said that), but I'm merely pointing out the problem. I'm not sure how you confused this with me providing the final say on the issue. I'm raising the issues to be considered, which is why I started my original post with "The real issue, in my opinion, is ..."

TL;DR.

If mIRC included the Aspell library (what Firefox and Pidgin IM use), then people who use Pidgin to IRC instead of mIRC couldn't use the excuse that "Pidgin is the only client with a spell checker."

Integration of Aspell takes about 5 to 7 minutes. It works with mIRC's existing RichEdit control (the place you type). Squiggly lines become automatic.

In an attempt to not let this thread totally spiral out of control, I thought I'd just make some general, on-topic points:

I don't really understand why you are so adamantly against the inclusion of a spell checker given that if you didn't want to use it yourself, you could simply disable it. Although I don't agree with every point you've made, the majority of them are true and not really being contested, so I'm not entirely sure why you keep listing off the flaws of spell checking. No one is claiming that spell checkers are perfect, but I would still find them useful even in their flawed state.

The point about it being time consuming and/or complicated to implement is certainly valid, and is the real crux of the discussion. I never really had any opinion on this point.

Firefox and Pidgin are both open source software. Can Aspell be used in a closed source application like mIRC without violating the license?

/help Acknowledgements

Firefox uses hunspell, not aspell, (as does Chrome, btw) and hunspell is used in proprietary software (according to the site), so it should be fine. It's GPLv2, which allows distributing a DLL as a separate binary entity so long as it wasn't modified, and the license of the library does not have to be applied to the entire program (mIRC).

Googling "hunspell richedit" (or aspell) shows that it's very much a nontrivial integration. That search returns very few resources and many people asking how to integrate the two with very little success. It might be possible, but it's not trivial. Definitely not 5-7 minutes. Otherwise you (or somebody) would have done it by now, instead of posting about the need to implement spell checking.

Frankly, if it really is that trivial to make, and is meant to be opt-in for the few who need it, this seems like DLL territory to me. As a DLL you'd be able to keep it up to date way quicker than mIRC could, which might be useful.

I've been using tinyspell (a spell checking software designed to work in every application) for a couple of days now, and I'v enjoyed every second of it. Frankly, I don't see why some of you are so adamant about keeping mIRC spell-checking free.
Adding missing words to the dictionary is only a matter of a couple of hours of chatting, and once spell checking is implemented it won't take long before people will start sharing their custom dictionaries to avoid having to add the missing words themselves, so that's one worry gone.
Another argument claims that people will just ignore the red squiggly underlines. As that might be true, people who will ignore it are probably the same people who will just turn off the spell checking feature all together. I mean, do you ignore chrome's spell checker? Probably not. If you care about spelling while chatting as much as you care about it while writing a comment on a forum, then I'm betting you won't ignore the red squiggly underlines. I certainly don't ignore tinyspell's tip boxes.
Sure, it won't stop people from mixing up "pic" and "pig", but that's just being petty... False positive/negatives exist in every software that has spell checking and it's just a matter of getting used to it and knowing how to spot the errors.

This approach could be added quickly but if you start to parse richtext that might lead to unexpected behaviour and will require additional coding.

If you've actually been following this thread you'll see a nice image that proves at least *I* have been ignoring it. I had no choice, because it was wrong. Is your spell checker always correct about the tips it pops up for you? Certainly not. So you will occasionally ignore your spell checker too.



The problem is that you're making a false assumption. Most people do not put as much effort into text on IRC as a forum. This is quite clearly evidenced by the lack of proper capitalization and punctuation over IRC vs. over a forum. The difference stems from the fact that text on IRC (typically) isn't published and archived on the internet for eternity, so the need to expend extra effort to make sure your temporal messages are perfect is usually reserved for people with OCD tendencies.


Finally I say this:

If there's a third party tool that works, why not suggest users download and install that rather than requiring this be re-implemented in mIRC? Then Khaled doesn't need to do anything, and you don't need to wait for this feature. You seem to already have what you want.

As a sidenote, the irony of your post is that you actually have a spelling error in your very first line. There is so much to be said about the "spell checkers are awesome" guy with a spelling error in the very same post, but I probably won't do it justice-- I'll just let it sit there for you to think about:

My English is not perfect and I'm aware of it. I'm not a native speaker as you have probably assumed and that's one of the reasons I want a spell checker in mIRC. Maybe spell checking is more of a feature that fits foreigners better. I'd probably turn off spell checking for my native language now that I think of it.

The software I mentioned is not free and has a lot of flaws/bugs (that wouldn't exist if mIRC implements spell checking natively) that I won't get into describing because it's not really relevant.

You seem to have a hard time understanding my point about false negatives. As I said, false negatives exist and you're right to ignore them. Mind you, ignoring false negatives and actual negatives are two very different things. When I asked if you ignore chrome's spell checker I meant ignoring actual negatives, not false negatives. Of course you ignore false negatives, you'd be a fool not to (why would I even suggest not ignoring false negatives?). Anyhow, some of these so called "false negatives" are not false at all since words like mIRC, aspell do not exist in the English dictionary (the same thing goes for the acronyms 'lol', 'IRC' and alike) - it seems that you have a rather subjective view on false positive. I've also mentioned in my previous post that it only takes a couple of hours of chatting adding those particular words into a custom dictionary that will validate these words.

I wasn't making a false assumption since I didn't say all (nor did I say "most", for that matter) people care about correct spelling while chatting as they do while writing a post on a forum, I said _some_ people do (notice the difference between some and all/most). There's also a huge difference between spelling and sentence formation/grammar (correct forming of sentences, punctuation, etc). I never said that people on IRC care about capitalisation or proper punctuation (although I'm sure _some_ of them do), I simply said some of them care about correct spelling of words. You could argue that spell checkers keep well formed sentences and capitalisation but that's simply not true in all cases, and it's certainly not the kind of spell checker that I think needs to be implemented in mIRC.

And by the way, the error I made was a grammatical error, not a spelling one, and it's far from ironic (see first paragraph) (Thanks for pointing it out to everyone by the way... That was nice of you). You had quite a few grammatical mistakes yourself, just so you'll know.

Really? So everything mIRC implements will be bug free? I understand there are integration problems with a product like tinyspell; but the bugs and flaws are likely the product of poor development, something even mIRC is not immune to. Perhaps try a better product. A quick google search shows that tinyspell is neither the only nor the most popular global spell checker for Windows. Some cost money, yes, but you're also not just investing in mIRC. Certainly there will be other programs that would benefit from global spell checking on your system too, if you really love spell checking. Also, a global spell checker would share the dictionary, so if you add lots of words, that's a big plus too.

FYI tinyspell has a free version according to their site.



And I had already mentioned my disagreement with this assertion in prior posts. Specifically, you're making a few assumptions: 1) the user will "add" a new word every time they come across one. My assertion is that most won't-- this goes back to my false positives fatigue UI issue. If you're presented with a red line every third word you write, eventually you stop reacting and start ignoring. My guess is the only ones who stick with this will be those with OCD behaviours. 2) You're assuming that eventually they will stop using jargon that is not in the dictionary (or that there is some asymptotic curve for the interval with which they get false positives). I take issue with this as well. Most people using IRC for technical discussions (which tends to be a large portion of people using IRC), have an uncountable arsenal of technical terms; the well does not run dry here. I can't even enumerate how many technical terms I personally know (and might use)-- adding them all would likely take way longer than a few hours of chatter. Finally, 3) you're assuming that they only have one dictionary of words to fill. Quite a number of users use IRC in multiple languages. Filling one dictionary full of jargon is one thing, but doing it in each language you speak over IRC? This has UI fatigue written all over it.



Then you're asking the wrong question. No, I won't ignore an actual error, but I wouldn't care that much if I had over IRC (where data is temporal in nature). I have made spelling errors on IRC before. I correct myself, or someone points it out to me, and life goes on. Eventually that data disappears into the backbuffer, only to be seen accidentally when perusing a log file for a URL that I forgot about. The issue is, though, that for every actual spelling error that a spell checker catches (I'm using Chrome as an example here, but this is technically a bad choice, since my text on IRC is a lot more mangled than it is here), there are about ~50 false positives/negatives that I have to "ignore" as well. And that is a very conservative estimate. My guess is this probably is true for many other users, but you can disagree with this statement if you'd like.

You can actually test this theory (and I'm tempted to, myself) by scouring your log files on an arbitrary large channel, and count the number of actual spelling errors, then run the text through a spell checker and see the total number of errors reported by the tool. My guess is number of legitimate errors will be a small value, even on the absolute scale, and the number of false positive/negatives will be orders of magnitude larger. This is why I question not the purpose, but the effectiveness of such a feature. Is it worth the effort to catch a handful of errors when there are third party tools for those who care enough or need it? I don't think the idea that "spell checkers will help people and therefore be worthwhile at any cost" should necessarily be taken at face value. Studies have shown that they don't make a difference for primary-school children, which is admittedly not the same demographic, but still somewhat enlightening. I'm not questioning that spell checkers sometimes work, I'm questioning the value-add of this feature; will it work often enough to be worth the effort put into integrating and maintaining it?



No problem. It's not every day the OP helps make my point for me. It was a spelling error, though. "I'v" is a typographical error-- your grammar was correct, you missed a letter-- your spell checker did not catch this because "v" is only one letter long (we discussed why single letters are exempt previously in the discussion). You'll notice that "I'vx" does properly show up as an error in Chrome. This goes back to my point about the ineffectiveness of spell checkers.

In general, the only people who actually care about spelling on IRC are the people typing. Rarely have I seen anyone complain about someone else's spelling unless it's so bad that you can't understand what's being said. If no one reading what you write cares, then the benefit is very minimal.

I care about spelling and I will point out spelling and gramatical errors to people when I see them on things that aren't temporary. I don't point out errors on IRC because, as argv0 stated, it is temporary and fixing errors has no value there. I also limit where I point out errors in other locations, such as forums. Although posts are more permanent, pointing out errors just derails the threads. Besides, posts are rarely significant enough that errors matter. It's not like a resume or essay where errors can cost you. The point is that IRC really isn't a medium where spelling and grammar matter.

Even though I care about spelling and grammar, I usually ignore the red squiggly lines from spell checkers. In most cases, they are wrong. I also rarely add words to the dictionary unless I get tired of seeing the red squiggly line on the same word over and over. I do, however, pay attention to the green squiggly lines for grammar because those are often right. A spelling error or typo now and then really isn't as bad as horrible grammar, imo. It's very easy to overlook a spelling error when reading something, but bad grammar sticks out like a sore thumb. Even so, I don't think we need a grammar checker on IRC either for the same reason that I don't see any significant benefit to a spell checker. It is temporary and *very* few people care if you type something incorrectly.

Here's an example of how bad spelling usually doesn't make something difficult to read...



Note that no such study was actually made and this is just a common example of how spelling errors don't necessarily make something difficult at all to read. Also note that not everyone can read this as quickly as a correctly spelled sentence and anyone who isn't a native speaker will probably have more difficulty. However, most native speakers can read it as well as a normal sentence.

The point being that spelling errors aren't the end of the world and that correcting them really only matters on something that is permanent, or at least semi-permanent. Other than logs that very few people read, IRC is not permanent at all. Grammar is usually more important, but it's still not important on IRC.

As I said, I actually do believe this is on Khaled's list of things to do. I don't think it's very high on the list, though. Anyone who really cares about spelling should look into 3rd party alternatives because it could be years before you get something added to mIRC.

Actually, I'm a much better speller now because of spell checks than I was in school, so this theory is incorrect. I learned over time how to get rid of the little red lines.

Auto-correct on the other hand is pure evil.

I'm certainly all for a spell checker in mIRC. I'll have to check out some of these scripts.

Proof by example much? One success story does not prove every case. It's also not necessarily even a success story, since this is not a proper controlled experiment. How can you really say you became a better speller because of spell check and not, say, just due to simple experience/practice of writing multiple papers over the years? You can't-- there's no control. This is exactly what conjecture is.

For years I've been a user of mIRC and one of the features I have always long for is a built-in spellchecker.

Through the years I've seen many scripts claiming to do such job but they all fail. I've also seen many threads with discussions about this matter and the one thing that ends all conversations is the Dictionaries. Well, why can't mIRC just come with a built-in spellchecker [spellcheck support] and we the end-user install Aspell or the dictionaries found at OpenOffice.org? What is the problem with this particular scenario?

A good example of a program that comes with spellchecker support but without the dictionaries is Pidgin [ http://pidgin.im/ ] If the end-user really cares for the spellchecker then it has to download and install the dictionaries either from OpenOffice.org or Aspell.

It isn't that people wont notice the red underline, it isn't that people in chats don't care; the fact is that throughout the years hundreds of people have requested and shown interest for this feature.

Please mIRC, please.

Hundreds of people wanting something is a small drop in the bucket considering the many thousands (hundreds of thousands probably) who use mIRC. There is far more support for emoticons (or images in general) than for spell checking and that's still not part of mIRC. Hundreds of people isn't much considering the number who use mIRC, so that's not really a great reason to add a feature.

If you read through the thread, I think you'll find that using external dictionaries is already what people are suggesting. No one is suggesting that mIRC create its own dictionaries. Or if they are, they're crazy. That would require way too much time for Khaled to maintain many dictionaries for many languages.

As I pointed out, Khaled most likely (this is a guess) has this on his To Do list. But don't expect it to happen anytime soon. For one, mIRC updates that include major features usually only come out about once a year. And spell checking is not a critical feature that would benefit the most people, so I am sure it's far down the list (just like emoticon support). I'm sure it will happen some day... but it probably won't be for years.

I haven't even seen "hundreds" of users ask about this feature. It seems you are exaggerating the issue. In fact, I just did a search on "spell" in the forums (over ALL forums) over the last 5 years, I only came up with 59 names-- note that not all of these names even support spellchecking (I remove the obvious ones), so it would probably be more like 50 names. 50 is a far cry from "hundreds", let alone even one hundred. And this is over 5 years.

As a paid user I'd like to say I'd love to see aspell & or hunspell support added as an option. X-chat has it, it would be great for mIRC too.
Anyone that dislikes spell checkers could just leave it turned off.