Forum Posts Feed
Text expander and clipboard enhancer tools
< Next Topic | Back to topic list | Previous Topic >
Pages: ‹ First < 6 7 8 9 10 11 12 13 14 > Last ›
Posted by Dr Andus
Oct 21, 2013 at 09:22 PM
jimspoon wrote:
>just looked at the webpage, $55.95 is pretty steep!
Then you’re in for a surprise, because the SmartComplete feature I was talking about is only available in the PRO version, which is USD139. 95 ;)
>Your description gave me an idea; perhaps it has already been
>implemented by somebody.
>A utility could monitor text that is input, and record commonly used
>phrases into a database. (No need to manually create a database of
>autotexts and abbreviations).
The AutoHotkey script I linked to before has such a “learning” feature and it’s free (I haven’t tried it). But obviously it would be much more of a DIY experience:
http://www.autohotkey.com/board/topic/49517-typingaid-v219d-word-autocompletion-utility/
As for why these text expansion software are so expensive? One reason might be that they’re mainly targeting medical doctors and business organisations who can afford them.
In the case of PhraseExpander, I’d say it’s in the premium segment of the market (and from what I’ve seen so far, probably the best), so it would be difficult for it to price itself lower than its main competitor, PhraseExpress. And as for the others, none of them seem to have the SmartComplete-type functionality, which is the only thing that really interests me.
I’d say these kind of tools are for people who are not touch-typists (or who suffer from RSI) and who type a lot of repetitive and long phrases all day long, use complicated terminology that are easy to misspell (medication, chemicals, strings of computer commands etc.). It’s probably the same people who also buy Dragon NaturallySpeaking (which is also not cheap).
But another way to look at it is that if this kind of text expansion software works for you, it would be *the* software that you use the most (besides your OS), because it literally monitors your every keystroke.
Posted by 22111
Oct 21, 2013 at 10:07 PM
You both present very interesting concepts here I fully agree with.
As said, typing the whole word is often faster than to identify the term within the drop-down box, then activate arrow keys, or (if they are numbered, which is the case in some such programs) press “1” or “2” or “3”, and this advice here:
“E.g. I can type “con” and the list “concept, conception, conceptual, conceptually” shows up in a drop-down box. Now I could use the arrow key to go to “conceptual”, or click on it with the mouse.”
is a good example of what I mean, since those terms are not “complicated” enough in order to justify your hand switching from keyboard to arrow keys, let alone to mouse, and back. (Btw, you bring a good example for my syllabus advice: ion, tual, tually…)
But of course, as soon as the tool in question comes with some AI (or at least brings my concept to work), things get VERY interesting:
“But with option 2, I could also type “conl” (first 3 + last letter of “conceptual”, and that word jumps to the top of the list, and then I just hit the confirmation key (Tab key for me).”
This is almost perfect, the only point of “criticism” I would have is, how to avoid the trigger key in such examples? You could do it by the negative: If you type the “l”, after the “con” will show the list, the tool “knows” you want “conceptual”, and if you had some other word in mind, you would type some # and then something else - I know this brings a lot of additional problems…
Also, WHEN will the drop-down list appear: For any terms the tool has in its big lists, or just for words you often type? Here, real AI would come into play: It’s imaginable that a text expander tool, just as a dictation tool, would be able to analyse your texts (even in groups), and then only presents you such “terms to choose from” for longer terms you really use (special vocabulary of your profession), and this would imply that it would not present any “improbable” term, since even READING this list within the drop-down frame takes much time.
Optimization is always possible: No presentation of anything else, but just 2 or 3 terms (or just 1), together with some very big “l” in the term so you know what the list will trigger then, and a very big “y” for the “conceptually” (and not an “ly” since the “l” would already trigger the “conceptual”, without extra trigger key).
If you want to write something else, why not enter a dot? Since we are in the “middle” of such a “presented list”, we can assume you are in the middle of a word, and not at the end of a sentence - such a system could be practical if it only showed you such lists for “special” words, and not 12 times for every line you type!
Further AI could go beyond analyzing your texts, from which analysis then it would constitute its “terms offerings” tailored to your texts, but it could go a step further and eliminate any “offered” where from the semantic context it could not apply: The tool could detect it’s a noun that’s needed here, not the respective verb, and if it was able to do so, the very SAME abbreviation could trigger the verb or the noun (and perhaps even without presenting a list to begin with).
Also, such more developed AI would be a thing from heaven for German users (I’m sure Mr. Bartels reads here, and if he can realize it, our money will ask to go for him !), since we have all those “flections” or whatever the linguist calls them, different word endings depending on all sorts of things… but which by AI would be detectable!
Just one example, it’s not “the”, it’s “der, die, das” for the article, “die” being female or plural. Now if you begin to type a word that is not female, but if also there is a “die” before it, any tool could/should “know” that you want to type the plural form of the word you’re typing, and in German, in many cases, that’s not some simple end-s, but all sorts of different things, so this “guessing from context before” could be extremely helpful.
A last word: Nobody says that memorizing abbreviations is sensible. It’s just that those drop-down lists take reading time, then take time for your reacting accordingly, so if somebody could optimize this concept, anybody would be happy with that!
Sorry, Jim, for replicating what you already said: “A utility could monitor text that is input, and record commonly used phrases into a database. (No need to manually create a database of autotexts and abbreviations).” And I should add, the tool should “judge” by frequency and by “complicatedness of typing it manually”, but in the end, that seems a complicated task to accomplish, and then, I would like to have my say about what I’m expected to type! (But this could be resolved by “smart global replacement”, meaning the tool would also check your corrections of what it wanted you to type for other instances of the same word components.)
“Then, while typing, when you want to insert any one of those commonly used phrases - you might type in a keystroke that signals the utility “i want to insert one of my commonly used phrases. use the following keystrokes to find that phrase for me.” I often use the forward slash “/” for that purpose.”
This means, special trigger keys for special lists, from which to choose then. You can perfect this system by installing a tree of such lists… (And then, some - very basic, easy - AI - would bring up your real standard phrases onto the very first page of this list collection.)
“Now just start typing parts of that phrase - with spaces to signify an AND condition.”
This is a brilliant concept which can be further developed: Whenever you type something devoid of sense (in the language the tools knows you are currently typing in), for example “as” in German, the tool knows the dot, or the space, or the comma, or many other such readily-available characters, cannot be a regular trigger key, and isn’t end of word or of sentence (part) either, but must be something special (again in several “lists”, depending wheter it’s a dot, space, comma…), and so it will “do something predictable” about it.
Btw, there is a medical text expansion program costing around 500 dollars if I remember well (so Bartels stuff is not THAT expensive really!) - unfortunately, you won’t get a trial without their knowing your name, address, tel number, etc., and they even phone you in order to check!
But it would certainly be of interest to know what really expensive text expanders might do “on top” of what we know from them.
Posted by Dr Andus
Oct 25, 2013 at 12:35 PM
Dr Andus wrote:
I’m still working up to my PhraseExpander review but I just wanted to
>point out that the latest version (3.9.6) released today has a new
>algorithm which optimises its “SmartComplete” feature (drop-down
>selection box at the cursor)
I’ve written a more detailed explanation of how the new PhraseExpander algorithm works:
http://drandus.wordpress.com/2013/10/25/phraseexpanders-new-algorithm/
The key point is that the SmartComplete box gives you a visual clue as to what shortcut to type in order to select the desired word, which is much more useful (to me at least) than having to memorise the shortcut, and it also obviates the need to use the arrow key.
In the meantime I also came across another software that has a popup box feature (though I haven’t tried it yet):
http://www.smartype.com/index.html
But as I said, the popup thing seems to be a premium feature, as this one is roughly in the same price band (USD 120-150).
Posted by Dr Andus
Nov 26, 2013 at 06:14 PM
Dr Andus wrote:
jimspoon wrote:
>>just looked at the webpage, $55.95 is pretty steep!
>
>Then you’re in for a surprise, because the SmartComplete feature I was
>talking about is only available in the PRO version, which is USD139. 95
>;)
PhraseExpander Pro is available at a 50% discount ($69.95) until Mon 2 December 2013 at 23:59 PST here:
http://www.phraseexpander.com/black-friday-discount-2013/
Posted by 22111
Nov 30, 2013 at 02:59 PM
Two things:
It’s 55,50 euro PLUS vat, and so we have got here another example of not-totally correct dollar-vs-euro conversion.
But I also would like to say that both “Dr Andus” ’ intervention at the developer and his blog article about the results, and the developer further developing his tools, are of admirable brilliance, and that the results cannot be underestimated; I think this has been a decisive step in the right direction.
Some remarks, all perfectly minor in comparison:
- I’m not happy with the tab key as the trigger key, should be any key, e.g. the one between my “jklöä” and my return key?
- I’m not happy with the pop-up window always appearing. I understand that it should appear when you cannot be sure yet, but in (almost) unequivocal cases, the tool should just type the full word, without any list popping up
- the same for the trigger key being needed
Now how could this be achieved? Perhaps this way:
We are in face of “learning software” - as said, the relevance of this cannot be underestimated. Now if the tool has any doubt since there are several possibilities left, with your typing, it will pop up the list, and in the form “Dr Andus” describes, meaning presenting it by presumed order of “chance” for the terms it deems possible here (this presumes some terms might not be yet within the respective “sub-vocabulary” yet), but if the tool has learned, from your previous use, that your key combination would indicate some of your “standard” terms, let’s say a word you will have selected at least twice from the list, even when technically there would be another alternative (but which you did not select once), the tool should simply enter this exact term, even at the risk of “guessing wrong”.
Of course, whenever such a guess is wrong, so that you must correct it (there should be, and probably is, a dedicated key combination to trigger this), the tool should then choose other “completing characters” for the lesser (or not-at-all) used words that theoretically might apply, and there should be “completion char” selection by frequency, and even by easiness of typing (and this from start on), which means that for frequently-used words (= for new users, the tool should analyse the texts you will have written without using it), the tool should select completion chars that will come “naturally”, on a standard keyboard (ultimately, even USA/GB/D etc. vs. F/E/I etc.), i.e. which can by typed by another finger, and/or by your other hand, and in the same line of thought, less frequent words (in YOUR texts) should ask for longer shortcuts than words you’ll be in need of again and again, even if they are “rarer” from a general-frequency statistical pov.
Once, I read somebody saying that this Nagarsoft product wasn’t but a paid replica of another, free program. It seems they have come a long way from that possible starting point, if it ever was true. Congrats!
Of course, any such tool (and of which in this case the possibilities are much broader than text expansion in AHK is) will not replace AHK (or AI or any other scripting language) for their non-text-expansion functionality and thus should work together with them without fault; for some cheap text expanders, this is not always the case.
But it seems that PhraseExpander is the top offering in text expansion today. I will continue to do my text expansion with AHK for the time being, but I’m eager to buy this commercial whenever it becomes near perfect, and chances are it will.