A simple reason why grammar-check is needed

[Vincent]

You ever write a 100+ document in Mellel, exported it to rtf, ran it through Word's grammar-checker, and found out that in various sleepless nights you managed to to write double-words like like this this one? I did, about 10 times. Not only did I have to run it through Word's moronic grammar-checker, but I also couldn't import it back into Mellel, as things (that appear in the new Beta) don't translate well there and back. So, while it is too late for me—many hours have already been spent in Word + actually reading what I wrote (haha)—I think that a "double-word" (let's not call it grammar) checker is a vital function needed in any long-doc writer.

Thank you in advance.

[petrus]

I have to respectfully disagree. I think a grammar checker is feature bloat. There is no substitute for careful proofreading, even though it takes time. Even spell-checking doesn't catch some errors. When I was using Microsoft Word, although it could be helpful at times (as in the double word instance mentioned), more often than not it ended up disagreeing with me, because grammar is such a complex system. Programming a word processor to check grammar is probably a lengthy, involved process. Furthermore, Mellel is used by so many people who work in so many languages, it seems unlikely that a grammar checker could be effectively implemented.

However, I do not know if there is a core grammar checker in Mac OS X, similar to the core spell checker. Perhaps that could be used, as Mellel uses other services. Another alternative could be to develop a module similar to the hyphenation module already in existence, that can be edited by users in many different localizations.

[ptram]

Hi,

Can't this one worki in parallel with Mellel?
http://linguisoft.com/gramprox.html

By looking for "grammar checker mac" on the internet, it seems to me that there are even some other programs of this kind.

Paolo

[Vincent]

@petrus: it's built into Leopard, which runs fine on my G4, Textedit is as fast as ever, and you can turn it off, so how is it bloat? I'm not asking the Redlers to build anything new here…

And even a proof-read can miss small things like to to and the the, nor is it, frankly, what I want to be proof-reading for.

[miles11]

You ever write a 100+ document in Mellel, exported it to rtf, ran it through Word's grammar-checker, and found out that in various sleepless nights you managed to to write double-words like like this this one? I did, about 10 times.

I agree with petrus, I don't think a grammar checker would be a practical addition to Mellel.

Long ago, I used a program called Grammarian. It was more of a PITA than of any practical use. And from what I've read, because of the complexity and flexibility involved, there is no grammar checker that can really be relied upon. One of the things Grammarian would do is bring up a separate window showing its opinion of a questionable phrase. The user then has to read the comments, think about this 'opinion' and make a choice. It doesn't take long to realize this could all be done better and faster without the program. As soon as its novelty wore off I gladly dumped the program.

You might have a look at Spell Catcher. It would have caught the double-words in your text (while you were still in Mellel). I've used it for years and still love the program. In addition to checking spelling, it includes a lot of other tools that can be useful to writers.

[Vincent]

Guys, with all due respect, I don't want to pay for a feature that already comes with a product that I paid for: Leopard. In addition to this, I think it's a safe bet that this core-feature will be included in all third-party word-processors sooner or later, making it a feature that people base part of their product choice on. I understand that grammar-checking is dumb, I would never want to rely on it 100%, but for things like double-words, I'm sorry, but that really is a time-saver. And when you have, as I said, 100+ pages to go through, you are not only prone to miss those small errors, but appreciate any time-saver you can get.

I can't wait to be finished with this thesis, I'm not planning to be a professional writer, and Mellel should respect me as that kind of customer. And, I would not suggest it, if it wasn't for the fact that Mellel uses only half of what is an optional (you can turn it off too!) feature provided by Mac OSX Leopard.

(If I sound cranky, it's because I spent half the night up finishing the damn thing. My apologies.)

[nicka]

Mellel's find function can fairly easily be persuaded to find repeated words.

Here's a recipe for the Find field:

Word-boundary element. Letter element set to 'once or more' and 'greedy', with brackets around it. Whitespace element set to 'once'. Backreference element, set to 'one' (this finds the same again as the thing in brackets before). Word boundary element.

I can explain in more detail if necessary.

It's not a good idea, by the way -- although it is possible -- to use this to Find and Replace All repeated words. I just ran this on my thesis and found about ten examples, only two of which were mistakes. It's better just to use this to find possible problems and use human intelligence to see whether they are legitimate repeated words or not.

[Vincent]

Interesting, I hadn't noticed that. But, if I understand it your, very technical, explanation correctly, you still have to enter the word to search for?

[nicka]

No. That's the point. The find elements I gave will find any repeated words.

[Vincent]

OK, maybe I'm doing something wrong. My search-box now looks like this, but doesn't seem to work.

Click for full size - Uploaded with plasq's Skitch

[nicka]

My fault. I should have explained more clearly that the brackets you need are the [square] ones you get by clicking on the bracket button just below the Find field.
The purpose of the brackets is to allow a kind of backreference within the Find-string. To find repeated words, you need to find a word, then find the same word right afterwards. The greedy-string-of-letters-element coming after a word-break finds a word. The backreference element after one space checks to see if there is the same word right after it.

Otherwise what you have looks fine, so fixing the brackets should do it, I think.

[Vincent]

Hey, that's great Nicka, thanks! I found four more repeated words, which Word's spell-checker and my proof-read missed!

I guess this sort of nullifies my request, though I still think that Mellel should have the same basic features that other Leopard apps have at their disposal.

[nicka]

It would at least make sense to ship Mellel with a few useful Find Actions like this repeated words one. Mellel doesn't really show off the wonderful features it has without the user putting in quite a bit of work. The lack of bundled Find actions is one example. Another one is all the wonderful OpenType stuff you can do -- but not with just a standard installation of the Mac OS and Mellel. I think they should licence and bundle a suitable font, or put a very prominent button within the application setup to download Cardo and install it, if they can legally do that.

When the Redlers have a moment I hope they sort some of this stuff out. I want Mellel to be a huge success, and little things like this that make it easier to see how powerful it is would surely help.

[petrus]

Yes, Mellel is extraordinary partly because of its ability to fully utilize OpenType typefaces, a fact that is not matched by any other word processor in the world. (Although Word 2008 advertises OpenType capability, I have heard that it's handling is still unacceptable, and Leopard's built-in OpenType support is very poor). I think that many people have not yet heard of the advantages of OpenType and Mellel's ability to access their features, such as true superscripts for footnotes and true small caps. I think that if the Redlers were to better advertise these capabilities, they would be even more successful.

[Roger]

Can you post some of your find actions? (Like this double word finder.) I was prompted to look at the Find feature in Mellel because of this discussion, and I have never seen such a powerful (and complex!) Find feature! I'm kind of lost in it. I'll have to read that section of the manual, but it would be nice to have a few examples to study and learn from.

Thanks,