Some enhancements/feature request that would be nice

[terceiro]

CatOne and kalisphoenix: wow. Thanks for answering. Those were great responses.

I’m still not convinced that using tags is useful when the information is already in the file (from outlines and such), but I can see how you might use it for both customer emails and submitted stories. Makes sense.

But at a certain level, I think that we’re all saying the same thing. We want a way to organize our stuff that’s faster and less intrusive than the current model. I want to be able to replicate a file in five locations at once, and with zero dragging. You want tagging. Everyone wants real smart folders. The fact that we have slightly differing proposed methods simply demonstrates that we’re different people. If the DEVON guys came up with a tagging solution that was functionally equivalent to my replicant-based system today, I’d swallow it without complaint. And if it were more efficient, I’d accept that.

To kalisphoenix’s original point, however, searching can indeed be a decent substitute to a tagging system. They’re different, but tagging is not inherently better than searching. Well, maybe for a straight string-match search, but DEVON uses some fuzzy searching that makes it much, much more robust than Spotlight (for instance).

I guess I resist tagging because it is so buzzwordy right now. In a couple of years the shine will wear off of tagging, and it will take it’s place as one of many options for organizing. But it is not, and will not be, the be-all, end-all of file organization. (And I don’t think that’s what either of you are saying.)

btw, kalisphoenix: 10 books at once? Focus, friend. Focus!

[terceiro]

kalisphoenix: I think we passed each other in posting. In your expanded example (Lance Throbbing? lol) I must admit, you made my head hurt, yes. I’m not sure my sphincter slammed shut, but I wasn’t thinking about my sphincter at the time.

I can only reply that my writing style and yours are very, very, very different. Even my (historical, partially fact-based) fiction and my scholarship would never need that kind of system. Not because it’s less intense (I presume) but because my brain doesn’t work like yours.

I use A-Z paper files and I use replicants in DEVONthink Pro in a pretty straightforward system. I have a group for each play by Shakespeare, and also a groups for themes: sex, power, monarchy. Files are replicated wherever appropriate. And more authors and themes and time periods. When I need information, I check the files. When I’m reading an article (or a note I’ve written), I use “See Also.” Simple? Perhaps. But it’s productive and efficient for me.

Yes, it could be done with tags and smart-folders, but there’s no compelling reason why a good tag-interface would be better than a good replicant-interface. They’re both options. I think I’m pretty close to orthodox in this. Your system, as far as I can tell, is pretty unique to you. In the long run, you might be forced to comply (or write your own). I don’t mean to imply any value judgement in that; it’s just economics.

Either way, I can’t get a straight answer about when 2.0 will be produced by the DEVON team. We might be talking about 2009 before we see any significant improvement in either tagging or replicant-management. Alas.

[ndouglas]

@terceiro

kalisphoenix: I think we passed each other in posting. In your expanded example (Lance Throbbing? lol) I must admit, you made my head hurt, yes. I’m not sure my sphincter slammed shut, but I wasn’t thinking about my sphincter at the time.

I can only reply that my writing style and yours are very, very, very different. Even my (historical, partially fact-based) fiction and my scholarship would never need that kind of system. Not because it’s less intense (I presume) but because my brain doesn’t work like yours.

I really need to start that thread in Usage Scenarios Yeah, I’ve arrived at that conclusion too – that I don’t plan my fiction like anyone else.

Your system, as far as I can tell, is pretty unique to you. In the long run, you might be forced to comply (or write your own). I don’t mean to imply any value judgement in that; it’s just economics.

Yeah, I’ve been wandering around between apps for a while. If it is economics, though, I believe DEVON could charge that extra $2-5 a copy to pay for their time developing it. I don’t think a single person would refuse to pay $2-5 a copy given the presence of an extra capability in addition to the capabilities it already possesses. $15? I’d still pay for it because, as I said, it’s crucial to me. I’d cheerfully pay $300 for a DTP with tagging and nested smart groups. I am staring at a credit card even as we speak.

It’s not a slap in the face to tell me that DEVON couldn’t afford to spend time on a feature that only I would use – even though that’s obviously not the case. It does miff me a bit, though, that Journler offers those exact capabilities I want and it’s free. And that I never would have registered DEVONthink if Journler was scalable enough to handle my project.

[CatOne]

@fbrzvnrnd

But what you mean for tags? Someting like columns in a sheet?
As I said before (but nobody answer me!) I have doubt about devonthink for non searchable documents. For example: how can I use devonthink to catalogue a handwritten papers? I can use the office version for scan the document, but I do not use the OCR and I need to insert data like the paper original position, what the document is talking about, who created the document, et ceterare. And I need those data for each document, I cannot create the same fields for each scanned document.
The only answer I thinked is to create a sheet with ony one record, copy to the folder I’m inserting the document and group it to the scanned document everytime I scan. But I was wondering if it is the best solution.

f.

No, I don’t mean something like columns in a sheet for tags. All I mean is, I want the ability to add some of my own words to a document.

DEVONthink already has functionality that is very much like tagging today – you can enter these words, free form, in the “info” area of a document.

So what I’m asking for, really, is CONVENIENCE. I don’t want to have to pop open the info panel to do this – I’d like to be able to do it inline (say, in the header, like you can put a bulleted list in an RTF document). Also, it would be nice if DEVONthink would keep track of these tags for me – and auto-complete my entries.

I gave some examples previously – I don’t really feel the need to clarify unless Bill or one of the DT engineers needs clarification. Also, if you’re having a tough time getting your head around this… just download Yojimbo and have a look. Seriously, it’s a free download for evaluation… and playing with it for 5 minutes should clear this up.

As far as tags and documents – I don’t care if the documents themselves are searchable or not. Tags can be used with either. For documents that are “opaque” (scans that have not been OCR’d, graphics, movies, whatever) obviously tags are the only thing searchable. But tags can also be useful for searchable documents – if it’s a PDF and I want to use the word “geflatzo” for a tag but it’s not in the PDF, what should I do? (again, I know I can use info for this… it’s a matter of convenience). For editable documents, sure I could stuff the tags right in there – but what if I forget and print it with the tags in there?

The point of tags is twofold:

1. Convenient way of adding user-specific metadata to documents… both to documents that are opaque and to those that are not opaque – the actual format of the document itself doesn’t matter (other than non-opaque ones are searchable)

2. Tags make it easy to create user-defined group (“smart groups”) and make stuff searchable/categorizable in multiple ways, in a manner that is a lot easier to deal with than replicants. Replicants are the manual way of doing smart groups, and are much less extensible (again, take the case of wanting to add another category – you must touch every document to add it to the new group, whereas if the tag was already there, you just have to create the smart group).

[Maria]

Hi,

a really intriguing thread… Basically, I think like terceiro that putting replicants into groups is the equivalent to tagging files.

For both systems you need preparation: Set up a tag system or set up a folder/group system.

about tagging in both systems with drag and drop:

“Tagging” in DT is extremely easy, but “the other way round” as compared to real tagging: cmd-opt-drag the file on a group in the floating group window. Of course: one drag, one tag. Could there be enhancements with marking the file and the appropriate groups and command “tag!”? This is already very nicely done with “See also”, where I can mark several groups in the results panel and click the move command, which creates replicants in those places.

“Real” tagging means drag tags to the file. I wonder whether this could be made possible on the GUI base in DT: dragging one or more groups from the group window on a document?

about typed tags
Typing is quicker than dragging. So typing tags into a tag field seems a quick solution, and one can create new tags on the fly. But how do you find spontaneously created tags later? What a about typing errors? There could be automatic completion as a solution.

Now I wonder whether DT could, in combination with the above mentioned UI issues, think of an option to create a list of groups by typing, where replicants of the file should go. (Although I personally think the solution with dragging those groups / tags in the group window is far better)

Checking, which tags are applied to a document

This is one of DT’s deficiencies: Checking all instances/replicants of a document (all tags applied to a document) is a mess, impossible in most cases. There should be a meta data field where the information is stored: “Replicants in groups:” and a list of those groups. Then we are almost closed to what tags are, but we do not destroy the DT philosophy.

As terceiro says, tags are a bit a fuzz at the moment, DT’s idea with groups is basic, and perhaps more persistent.

tags in comment field

I did this for a while, but soon I was convinced that the replicant-group system was meant for tagging.

I would like to hear from the developers what they think about this thread, it seems to be crucial for the acceptance of DT as a whole.

Best,
Maria

[talazem]

On the side of tagging, which has been mentioned in these forums before: a major benefit of tagging (as commonly understood, not through replicants) is that if (a) you ever export your work, or (b) your replicants become corrupted (which happened to me in DT before), you can easily recreate your “group/folder” structure through smart folders and the tags. I like replicants. But I also see a lot of wisdom in organized tagging.

[fbrzvnrnd]

@CatOne

[

2. Tags make it easy to create user-defined group (“smart groups”) and make stuff searchable/categorizable in multiple ways, in a manner that is a lot easier to deal with than replicants. Replicants are the manual way of doing smart groups, and are much less extensible (again, take the case of wanting to add another category – you must touch every document to add it to the new group, whereas if the tag was already there, you just have to create the smart group).

So for tags you mean tags… sorry for my silly question but in those days I’m jumping from a computer to another, so sometimes I lost the focus of this thread.
Anyway, the tag option is not a good option for my point of view. I understand your problem, but for a complex cataloguing of ‘opaque’ document, tags have some contraindication:
The first one is that tags are free, so I can make easily error forgiving a ‘field’ in comment. For example: some years ago I begin to list a large paper archive, using x-world (a object oriented database). Every time I got a document to list, I need to identify it and add a object with some field inside (the objects were stored in a model, with all the fields ready). Every kind of document has different kind of fields, and the kinds of objects was more than 60 different kinds… so, I can not remember for each kind of dcument I meet, the exact tags to add every time! Another problem with tags: if you put tags in 100 different document and after you did this, you understand you need another tag with a new value, you have to edit all the 100 documents and add the tags… With fields and model you simply change the model you used and all the document now have the new field. The same for delete, or modify the kind of a tag.
I see reading the forum is that DT is a versatile program, so all the people want it similar to their demand.
So I can understand your problem and your ask for a tagging system, but for my demand, DTPO is lacking in his database option for opaque document. Keep in order searchable documents is a good thing, but a database need something more to be used in professional use for archiving ‘opaque’ document. And it is strange to me that a powerful program like DTPO do not feel the necessity to query his datas in different ways. Even if I use the sheet to archive datas for the scanned document, when my boss ask me ‘hey, I need a list of all the legal transaction written by Smith Notary’, well I simply can not have it using DTPO 'cause I can not easly query sheet using column as argoment.

So, I feel the same demand you feel, but for another problem and asking for a different solution.

Sorry for my english and thanks for your answer.

f.

[ndouglas]

@fbrzvnrnd

Every kind of document has different kind of fields, and the kinds of objects was more than 60 different kinds… so, I can not remember for each kind of dcument I meet, the exact tags to add every time!

There could be something like a “Class” (or a “Category” or a “Type” – doesn’t matter). This Class would be a group of tags under a name. Click the Class name and it applies the specified tag to the selected document(s). This doesn’t have to be in DT – it could be a simple AppleScript. You would fill out the tags yourself (this is just a template of sorts). Problem solved and it requires no extra work from DEVONdevelopers.

Another problem with tags: if you put tags in 100 different document and after you did this, you understand you need another tag with a new value, you have to edit all the 100 documents and add the tags… With fields and model you simply change the model you used and all the document now have the new field. The same for delete, or modify the kind of a tag.

You could do the same thing with tags using a template-inheritance scheme. Dig Wikipedia – if you put a [[Category]] tag into a template (for instance, the tables that a movie’s information is displayed in), then all of the articles that include that template inherit its Category tags.

Which I’d like.

But I don’t want to overload the developers with requests, especially when they’re unnecessary That’s a v3.0 feature.

With the evolved tagging system I’d like to see, your problem would simply never exist. I’d select those 100 documents by whatever tag they have in common and then add the new tags. And edit the “Class” definition in my AppleScript or my XML config file for the Class definition, or whatever, so that newly-created documents get the correct tags. Just a few seconds.

[fbrzvnrnd]

I understand, so your solution could be good. But don’t you feel the tag system like something outside the devonthink philosophy? Everyhing in devonthink is an object, from the record to the scanned document, and tags seems to me something outside this kind to represent the database. And I feel the curve of learning asked to to a user to use tags and sacript is more than a system based on icon and columns.

Oblivious it is a only my point of view

f.

[CatOne]

@fbrzvnrnd

I understand, so your solution could be good. But don’t you feel the tag system like something outside the devonthink philosophy? Everyhing in devonthink is an object, from the record to the scanned document, and tags seems to me something outside this kind to represent the database. And I feel the curve of learning asked to to a user to use tags and sacript is more than a system based on icon and columns.

Why?

Tags are just text. A good tag implementation has a list of tags that you can apply to any document or object in the database. Type them in (have it autocomplete like mail.app does on an address), or select them from a list and drag them onto the document. The implementation of tags is just a list.

I don’t really understand all the obfuscation and complexity that people are bringing into this. Seriously, if you’re confused about tags, go download Yojimbo 1.3 and play with it for 20 minutes – you’ll see how it can be useful. If you don’t understand tags, quit throwing up strange use cases to make them complicated – that’s not what I’m asking. Yojimbo, Aperture, and iPhoto even have decent examples of what tags can do, and how they can make search a lot easier in cases where you want to use tags. Of course, tags can be optional – just like adding info to the “info” field is today.

[sgmiller]

I guess nobody was too impressed with my brilliant distinction between Scrapbookers and Researchers. So I’'ll try one more time.

I think the central disagreement in this thread over the utility of tagging (whatever the implementation) is being missed which is that there is a crucial distinction between filing and organizing. There is no question that tags could be very useful in filing. As somebody asked, what is the point of filing if you have to search 90% of the time to find what you want? Exactly! The original computer concept of replicating the paper folder missed the point that a computer can store things in more than one place at the same time (Microsoft is still trying to actualize this over 25 years later with NTFS!).

Relicants/duplicates are a clumsy way of trying to file in more than one place whereas tags are the elegant solution. As I have said before, I don’t use DT as a file cabinet (Scrapbook) so tags are not useful to me. I do wish I had started tagging my document archive, but it was started years before and I feel like it is too late, so I just dump everything into an archive and rely on searching to pull things out (see my posts on FoxTrot). On the other hand, I do use DT folders as organizing tools to establish a structure for report writing which is altogether a different matter from filing. I don’t need to search DT all that much, only to find the occasional document I am looking for.

Bottom line, I see where tagging could be very useful for Filers/Scrapbookers and DT columns seem like a natural for this by adding some kind of self-completing column entries. In addition, it would be nice if DT would improve the search/query function to allow for complex queries including columns. For example:

Find all records containing “John” and date: between x and y

It seems silly to have what amounts to database fields built-in and not be able to construct complex queries using those fields. As I said, I would not use this function as I currently work with DT but I can completely see why Filers/Scrapbookers would need it.

[CatOne]

Understand, sgmiller, and I agree.

I’m not an author, and I bailed from the system after I got my masters degree, so I’m not a researcher either. So I do fit the mold of “scrapbooker” as you have stated… but really I primarily do use search. The reason I would like tags is because tags can help guide search results. I can assign keywords for documents based on what makes sense to me, and based on what categories I want to pre-define.

Now it’s true that for the documents I already have in DEVONthink (roughly 7000 at present – about 1000 are RTFs I’ve written as notes, and 6000 are web archives, PDFs, etc.) it would be a real pain to go and tag every one. But that may not be necessary… and in many cases I could reduce the number of folders I currently have should I be able to nicely tag in the future… because right now I have a mix of folders + “big junk dumping grounds” that is half-way organized because sometimes trawling through folders fails to find what I want.

[sgmiller]

I think people who primarily search are Scrapbookers by definition.

Some of this discussion reminds me of a product that started out in MS DOS called AskSam. Maybe some on this forum know what I am talking about. I believe it was the first “freeform database” that allowed you to store all kinds of documents as individual records and then, if desired, to create structure later on by adding custom fields. It also had excellent search functions. This was very much a Scrapbook application, but the devs tried to add organizing functions in the form of crude folders which were essentially tied to the fields, sort of a folder/tagging system if that makes any sense.

I stopped using AskSam precisely because I am not a Scrapbooker and need organizing (Research) functions which is why I switched to their SurfSaver product and later to PC apps such as NetSnippets which allow you to quickly grab online and other content and organize it into a folder-based outline.

Interestingly, at least to me, is that DT has the basic elements of both NetSnippets and AskSam in that the columns are a bit like AskSam fields, albeit not as customizable nor as flexible in constructing queries. A little bit of thought and work though and this could be greatly improved for you taggers out there.

[sgmiller]

I should also add that I wonder if Search/Scrapbooking is the best use for DT. If I was doing purely that, I would prefer to store my material in normal folders and rely on the excellent FoxTrot search program to do the searching. I am not sure how DT improves on that unless you make use of the AI functions of course as well as the highly flexible folder system. Of course, the database is also far more transportable than many thousands of individual files if that is a concern.

As I have said several times, I do my Research/Organizing in DT but I do my Scrapbook/Searching using standard files and FoxTrot. I suppose that would be cumbersome for those who mostly rely on their own files for their work instead of constantly having to bring in and process new material as I do.

Anyway, it is just so interesting to see how many different ways there are to do “Knowledge Work” and I am sure this is just the beginning of a new science of some kind!

[fbrzvnrnd]

@CatOne

[

Why?

Tags are just text.

I’m not against tags. I can learn tags and use in a document, but I’m thinking to people working here. Tags are a textual way in a Os that uses a icon and mouse way. Some people here could never learn the use of tags. There is a folder, a menu, a icon, a field? Ok. I have to write something? Not. Wrong typed text, syntax error, tags forgotten or changed from here to there… tags have really a different learning curve for a non-computer-lover user.

For a personal use it is ok, but for a “masses” uses I do not think could be a solution.

f.

[fbrzvnrnd]

@sgmiller

In addition, it would be nice if DT would improve the search/query function to allow for complex queries including columns. For example:

Find all records containing “John” and date: between x and y

It seems silly to have what amounts to database fields built-in and not be able to construct complex queries using those fields. As I said, I would not use this function as I currently work with DT but I can completely see why Filers/Scrapbookers would need it.

I agree. As I said before DTPO lacks of something to query datas. And It is not a trivial problem, cause DTPO could not build relations. See this for example:

It is a simple database I’m making to list my books, stuff about the books I read, review etc. As you can say I have done a table with data about the books, and a folder for each book containing a replica of the record, the review, wikipedia info about the author, et ceterae.
Where is the problem? Well, if DTOP will have smart folder, or somethink like a query, this database fails, 'cause there is not a relation between the record of the book, the group of the book and the review for example. I mean: there is a relation, all the items are inside the same group, but there is not a way to formalize this relation for a smart folder or a query. I can not make a smart folder grouping file for author, or for classification place: i have those info in records, but I can not extend those info to other items in same group.
A way could be let grops have columns, like x-world did:

So the physical presence of an object inside a group become a implicit relationship.

Uh? Someone said dtp 4.0?

f.
[/quote]
[/url]

[ndouglas]

@fbrzvnrnd

I’m not against tags. I can learn tags and use in a document, but I’m thinking to people working here. Tags are a textual way in a Os that uses a icon and mouse way. Some people here could never learn the use of tags. There is a folder, a menu, a icon, a field? Ok. I have to write something? Not. Wrong typed text, syntax error, tags forgotten or changed from here to there… tags have really a different learning curve for a non-computer-lover user.

For a personal use it is ok, but for a “masses” uses I do not think could be a solution.

f.

DEVONthink Pro isn’t really a “masses” application. Tags are on just about every blog in existence, so would you really want to dumb down DT for people who aren’t capable of writing a blog correctly? How many of those people are going to be dumping into DT the number of documents required for the AI to be effective? How many of them would be cheerfully creating effective applescript smartgroups under the current system?

Tagging is an immensely inclusive (not exclusive) addition.

Besides, I can type a hell of a lot faster than I can point-and-click. If you want to make it even faster and less prone to errors, that’s why $DEITY provided the miracle of auto-completion.

[fbrzvnrnd]

@kalisphoenix

DEVONthink Pro isn’t really a “masses” application. Tags are on just about every blog in existence, so would you really want to dumb down DT for people who aren’t capable of writing a blog correctly?

Probably we are working with different kind of user. I work in a religious Insitute and here people uses the computer for normal office work: word, excel, someone make some simple internal dtp using pagemaker and photoshop, and oblivious bookkeeping with ad-hoc programs. This people could use this kind of programs 'cause are quite similar in GUI and have a low learning curve to be used. Internet is used only for some google research, some email and I think nobody here know what a blog is.

A priest ask me some times ago about a work I know he could do with DTPO. But I’m sure tags (or applescripts) were outside his reach.

I repeat I’m not agaist the tags, but I see tags like a additonal tool and not like a good integrated solution for extend the (actualy not much) tools for database istances.

f.

[ndouglas]

I just imagined my mom trying to use DEVONthink and recoiled in horror. I pointed a friend of mine at an open-source Personal Info Manager, and he’s about as computer-literate as the average kid my age, and he’s having a lot of trouble figuring it out. I can’t imagine the average clergyman or clergywoman having enough time and energy and computer knowhow to use DTP effectively.

There are much simpler applications out there for that sort of person. And free simpler applications, too. I view DTP as the Linux of PIMs: it’s staggeringly powerful but requires a good amount of work from the user. And I like it that way

But my point is that DTP is powerful and complicated. And other PIMs are simple and easy-to-learn. And most of them use tags in one way or another.

[Maria]

@kalisphoenix

…But my point is that DTP is powerful and complicated. And other PIMs are simple and easy-to-learn. And most of them use tags in one way or another. …

Good point.

So, what do you think about the idea in my earlier post: adding a tab to the info panel that shows the groups, the document refers to? This is what tags are. And you have the groups hierarchy included – tags in many applications loose their hierarchical context.

As for typing: I believe you are a fast typer, I am myself, even when typing German blind on a Japanese keyboard with US Layout. But I would not claim to tag a document with several tags quicker than I click on the respective groups in the group window.

I really do not understand the difference that is made between the concept of tags and groups, it seems superficial to me, the backbone is the same: it is a classification.

Looking forward for answers…

Maria