upcoming keywords pane

[wordmuse]

From the roadmap I see that a Keywords Pane is in future plans. I'd like to ask that you take a look at Evernote 3's use of KWs. Nearly perfect for my use. What I'd like to see in URP3's KW pane is:

1 - the ability to assign multiple KWs (already exists, I know)
2 - the ability to drag an item to a key word in the keyword pane and have it thus associated.
3 - the ability to see at a glance all the KWs for an item.
4 - the ability to click an item in the KW pane and have all items with that KW show up in the related window.
5 - the ability to delete KWs from the KW pane and have all items lose those KWs.
6 - the ability to select more than one KW in the KW pane and at my option have it serve as an AND or an OR.
7 - have a NOT function that I can toggle for any KW.
8 - have a query box where I can enter tags boolean style (nested, etc.)

I think this would make my day. Thank you very much. :)

Regards,
Bal

[critStock]

I'm glad to see that the keywords pain functionality is high on the list in the roadmap!

I wholeheartedly support wordmuse's suggestions, esp. #s 4, 6 and 7.

I would add one:
make it possible to arrange keywords in a tree hierarchy; that way keywords can be used as a "knowledge map."

Loving 4.1, esp. the "enhanced full-text search." That's a real winner!

Cheers,
critStock

[vogelap]

This would be an awesome addition. Please implement! I am a "guru" user of EverNote (2.2.1), and would be happy to provide opinions on the KW implementation in UR.

[vogelap]

Would love to discuss this feature!

[TMF]

My simple dream is to have a global keywords cloud pane which lists all the kw in the db and clicking on any kw shows the list of items assigned.

[critStock]

Hi, keyword watchers,

Personally, I've never understood the utility of tag clouds. They seem to allow only one dimension of organization--frequency. And if I have a lot of tags, it becomes difficult to find what I'm looking for in the cloud. When I'm looking for info relevant to tag x and tag y, frequency is irrelevant. Clouds look nice, but what am I missing that makes them actually useful?

What I am looking for is the ability to organize info items--therefore, a tag tree is what appeals to me most, where I can navigate around via relationships between tags. While I'm at it, if there is a tag tree, it would be great if tags could be cloned, the way info items can--this would allow me to create more complex relationships among tags.

There is an interesting, if incomplete, implementation in Surfulater. With any given tag selected, subitems are autoamatically created corresponding to other tags that appear on items in conjunction with the given tag. It's like an AND search, but unlike the AND search, it shows you all the tags found in conjunction with the selected one. The branch are iterative, i.e., you can keep drilling down to get new combinations. (Not sure I explained this well!)

I'm certainly looking forward to seeing what the folks at Kinook come up with!

Best wishes,
David

[TMF]

Originally posted by critStock
Personally, I've never understood the utility of tag clouds. They seem to allow only one dimension of organization--frequency. And if I have a lot of tags, it becomes difficult to find what I'm looking for in the cloud.
Absolutely.

My use for them is with small number of tags for fast navigation.

(I know I could use Search items for that, but I find search items annoying visually, and I prefer small pane where you have more kw on one line over a long tree).

[wordmuse]

Again - as I stated in the original post for this thread, I think that the way Evernote makes use of tags is excellent, both in functionality and aesthetics.

Not saying that URP should duplicate it; only that lessons & ideas might be drawn from it.

- Bal

[critStock]

@TMF--I get it. I can see how this would be ideal if you have a small number of tags and don't need hierarchy. For me, this might work for a subset of tags, but if it were the only view available it wouldn't.

@wordmuse: I'm curious whether you're speaking of EN 3.x or 2.2. A lot of folks have rejected 3.x to stick with 2.2. I'm not one of them, although I understand their point of view. Personally, I find the latest EN incarnation of tags quite good, although I would appreciate a more intuitive and powerful way to employ Boolean arguments. I feel like tagging implementations are still very much in flux across the software horizon. Tabbles, a file-tagging program, has a very ingenious graphical interface for Boolean logic. I can't imagine UR going this graphical, but perhaps it could be implemented in more tree-like mode, as in Surfulater, e.g.

Cheers to all,
dp

[wordmuse]

@critStock - I refer to the implementation done in EN 3.

And while I do enjoy graphical interfaces, I find that they can become cluttered really fast. So unless they are done with a lot of attention to usability built in, it can get old really quick.

For example, I can't really imagine how I'd use The Brain do do more than a fraction of what I can do with the Data Explorer Pane, though I'm open to learning. But I don't intuit a lot of value in such an implementation.

I'll take a look at Tabbles.

- Bal

[critStock]

@wordmuse: Yeah, the graphical interface thing is a mixed bag, to be sure. As I mentioned in an earlier post, I don't find the tag cloud model very useful beyond a narrow scope. I was using "graphical" somewhat loosely. What I'm really looking for is an interface that can provide me with a lot of information at one time. Not ALL the information, of course. And herein lies the detail devil. What do I want to see in a given "paper view" of my data? What are my options for determining the criteria by which my chosen view is narrowed down to my current view? That's what I'm really after, is an "elegant flexibility." If you do check out Tabbles, be sure to read through their documentation. The system is a little opaque at first, but if you spend a little time acquainting yourself with their way of doing things, I think you'll appreciate the power and flexibility of their approach. (Assuming, of course, that their goals correspond to what you need!) I've also found them amazingly responsive to feedback and support requests. BTW, Tabbles can also tag web pages, and recently they've added a very innovative and clever system for sharing/syncing tags between one user's pcs or between multiple users. It's definitely one of the coolest applications I've come across in years. If you get interested, I'll see you over in the Tabbles forums!

That was a bit of a digression, so just to get back to keywords (tags) in UR, a few features strike me as key:
(1) being able to access keyword functions for a given item quickly and with flexibility (e.g., can I quickly and easily add multiple tags to the current item?; can I get autofill for existing tags?);
(2) being able to explore items using a tree-like map of contingencies/logical arguments (tag=x + tag=y, or tag=x but tag=z)?: and
(3) being able to see and manipulate tags when viewing items as one does in the course of using UR (again, Evernote 3.x does this pretty well: you see the tags as little ovals with the tag text inside, and there is an "x" you can click to remove the tag, etc.

Cheers,
David

[critStock]

P.S. I agree about Personal Brain. The idea is attractive, but it doesn't really "answer" the kind of "queries" I typically ask of my data.

[vogelap]

@wordmuse & @critStock: I use EverNote 2.2.x every single day and absolutely LOVE it's implementation of the tag tree. (@critStock: EverNote 2.2.x allowed for creating "shortcuts" to tags within the tree, so you could bury the tag in a sub-sub-sub (etc) folder and create a shortcut to it on your top level.

I love that I can style (bold, italics, color, etc) my tag names, and that I can automatically assign tags based on keywords. AND that EverNote visually differentiated between automatically-assigned tags and manually assigned.

If URp were to improve upon that already-solid base, that would be amazing.

[critStock]

Those are great tagging features in EN 2.2. Yet another reason to mourn the direction EN has chosen to take....

[vogelap]

Originally posted by critStock
Those are great tagging features in EN 2.2. Yet another reason to mourn the direction EN has chosen to take....
I mourn it every day, my friend. Every day... At least EN2.2 still works for now (even under Windows 7).

I sure hope URp can take over those amazing features, and continue to grow them.

[wordmuse]

When EN evolved to ver3, I didn't really pay attention to the things lost but simply went with the changes. I guess I missed the features in ver2 that you mention, and I agree that they would be neat to have - in any program, including URP - that makes use of tags.

Especially if one distinguishes tags and keywords and/or enables rules that generate tags per text and attributes.

- Bal

[vogelap]

Originally posted by wordmuse
When EN evolved to ver3, I didn't really pay attention to the things lost but simply went with the changes. I guess I missed the features in ver2 that you mention, and I agree that they would be neat to have - in any program, including URP - that makes use of tags.

Especially if one distinguishes tags and keywords and/or enables rules that generate tags per text and attributes.
I sure hope that Kinook is taking note... I'd be happy to discuss suggestions for the tag implementation in URp! Tags are vitally important to me!

[quant]

Originally posted by wordmuse

Especially if one distinguishes tags and keywords ...

aren't they the same thing?

[wordmuse]

Originally posted by quant
aren't they the same thing?

Not in how I think of them, though they are clearly similar.

I think of keywords as being inline with the text, while tags, ideally, should be available as metadata with respect to the text. Keywords could be discoverable, e.g., if over time, certain words and phrases become statistically dominant, they would be keywords, whether I intended or recognized them as such at the outset.

Tags would always be explicitly created. Indeed, one might find it desirable to explore keywords to see what the tags might be profitable to use.

Does this make any kind of sense, or am I kind of off in my own little world here?

- Bal

[vogelap]

Originally posted by wordmuse
Does this make any kind of sense, or am I kind of off in my own little world here?
That is actually a very good definition! Thanks.

[quant]

Originally posted by wordmuse
Tags would always be explicitly created. Indeed, one might find it desirable to explore keywords to see what the tags might be profitable to use.
hmmm, one can already see Relevance attribute (which I guess is just a counter). I don't see for example a point in adding a Tag equal to keyword that is "statistically" dominant, because I can already find it by having a look at Relevance attribute. Can you provide some specific example where the advantage of having both Tags and keywords would be beneficial?

[wordmuse]

Just revisited how LinkedIn handles tags, and once again, I find myself impressed. It's not as robust as I would like, but it is so smooth and fast. And while they differentiate "tags" from, say, "locations," "companies," and "industries," all of these feel like tags and they are pulled automatically from the contacts' data, and then they are listed along the left side.

I recommend that URP include something like this in the upcoming tags pane if at all possible.

- Bal

[wordmuse]

07-22-2008, 11:14 PM
wordmuse wordmuse is online now
Registered User

upcoming keywords pane
From the roadmap I see that a Keywords Pane is in future plans. I'd like to ask that you take a look at Evernote 3's use of KWs. Nearly perfect for my use. What I'd like to see in URP3's KW pane is:

I can't believe I made that post almost 4 years ago. It was high on the road map.
http://www.kinook.com/Forum/showthread.php?s=&threadid=3204
Enhanced calendar pane
Keywords pane / tagging improvements
Add columns to Data Explorer pane

An enhanced keywords pane with tagging improvements was probably THE feature I was most looking forward to. I checked back at kinook.com/Forum frequently over this period, looking for this. Had Kinook not put it high on the roadmap, I wouldn't be caring as much, but what a letdown. Now I know what an "entitlement mentality" feels like from the inside. After all, Kinook had no contractual requirement to provide it. This is just mismanaged expectations. And, of course, it remains a 1st world problem. But the kid in me is disappointed, and the parent in me is saying, "There there, why don't you learn to be a developer or go see what else is out there instead of complaining?"

Well - I'm not going to become a developer, but the seeing what else is out there? Good idea...