This release is amazing, and also gives you *a full screen mode* to
surround yourself with the beauty of your own words and nothing else.
Even browser pages open FULL SCREEN if you want 'em to, too.

Its beautiful, to say the least.

The tutorial they have, thankfully, will clue me in on the nuances.
I've loved DEVONthink but I never really know just exactly HOW I should
best use it to leverage all the things it has - that is, for what its
worth, the only reason I don't use DEVONagent.  Bad documentation and
no real examples of why I'd want to use it.

I want to use DEVONagent because I'm sure its great, but c'mon.  I
can't wrap my thick head around it for the life of me.

Now I just need to decide if Pro is worth money or just fun.

I share you frustration with DEVONagent, however. I wish someone would
say, "Check this out! I used D.A. and was able to accomplish a) more
than every before and b) previously impossible things! Look at how I
did it!"

The team at DEVON are clearly a bunch of architypical programmer types
who can code amazing things, but can't understand that even amazing
things don't sell themselves.

[1] http://www.stevenberlinjohnson.com/movabletype/archives/000230.html

--jw

On 7/28/05, emory <emo...@gmail.com> wrote:

I need clear-cut examples of how someone uses it to their advantage,
preferably someone in an academia/university setting that does research.

It seems like one of those apps--like Emacs, Quicksilver, or Tinderbox,
I suppose--that most rewards the dedicated user. A casual user might
not benefit from using it for only _some_ info purposes, but, I can
tell you, I've seen some amazing stuff people can do with loyal use of
the app (mostly just over-the-shoulder demos).

I think in some ways I do a _really_ dumbed-down version of what SBJ
does, but with del.icio.us. When I'm researching a project, I can use
the del.icio.us tags and descriptions to herd relevant pages into their
own little pen and then make them easier to pull out later. It's not
nearly as semantically promising as Devonthink, of course.

Anyhow, I'd love to see something like a screencast of a Devonthink
jedi at work. I'm a big conceptual fan of the program, but I too would
enjoy a live demo highlighting the new features working
together--especially the "reward" moment when all the grunt work pays
off.

Been using DevonThink for a bit and man, it's well worth it.

The one issue I'm having, however, is that while I can store items  
(grouping, tagging, etc.) and search on them quite easily, would be  
great to be able to build some kind of personal taxonomy that I can  
browse.

Am worried that as I load DevonThink up with wonderful data that I'm  
forgetting about stuff I have stored. Now, I might come across it in  
a search, but you FIRST have to know what you're searching for. I  
find personal taxonomies (based on tags) great, because it provides a  
view whereby you get a list of the topics you've got data on and the  
volume of one over another.

Anyone know if this sort of thing is possible with DevonThink?

Oh, and one other item: The service they've got included that allows  
you to grab and store copy from ANYTHING using CMD + SHFT + ) is  
awesome.

Would love to hear how folks are using the app...

On Jul 29, 2005, at 8:35 AM, Merlin Mann wrote:

You mentioned tagging.  You can use 'labels' on things in DEVONthink,
but I think what bears more than a little interest is the Wiki-like
features they've been adding the last few months.

You could perhaps use Wiki-Like Tags to get back and forth between
documents that reference [[Tag]].  I'll experiement a bit.

I'm really enjoying the multiple databases now - I can put several gigs
of religious texts into their own database and build massive
correlations between items.

It has been proving to be very useful when researching various topics.

I have also been doing weird things like shoving all security
advisories into a database, and mining things that way.  It is a real
piece of work.

Ah...

Multiple databases. Hadn't thought of this. I'd figured on just  
adding tons of stuff in the single database, with filed grouped  
together. Your idea makes upgrading sound particularly interesting.

As for tagging, I'm handling it different than labeling. With  
labeling, I've got color codes for items to read and items read. Also  
have a label for items that I've actually highlighted content in or  
somehow otherwise annotated. That is so far helpful for me, as I only  
highlight or mark up some documents, not all.

Also, if you do CMD+I on a document, there's an info panel that comes  
up. It has a comment field and I can add tags there.

Again, having some kind of personal taxonomy list (similar to what's  
used on Flickr for tags) would be great.

As for multiple databases, when you're using the DevonThink service  
to copy content from one app to a DevonThink page (rich text or  
ascii), how do you specify the DB if you've got more than one? Do  
they allow for that? Or, can you shift files from one DB to another  
from a right-click option?

And wouldn't it be great if DevonThink had del.icio.us integration?

This app isn't 100% killer at this point, but it does 75% of the  
heavy lifting I'm looking for in a PIM and points clearly to the 25%  
that's missing that would make it even more useful.

/Anthony

On Jul 29, 2005, at 10:31 AM, emory wrote:

Either approach requires

2 strategies for building a "personal taxonomy" come to mind:

1.  Borrow
2.  Roll your own, a., from the beginning or, b., incrementally (as needed)

1.  Sources for borrowing such a personal taxonomy could include systems
such as the Library of Congress Subject headings for most of the deep stuff.
I'm sure you could find plenty of generic stuff for household management and
for prototypical office things or almost any common business or personal
role you might have.  There is a chapter in "Personal Documentation for
Professionals: Means and Methods", by Vladimir Stibic, 1980 on "Document
Description" that covers the matter in all the detail you are likely to be
able to find anywhere.  See also "Tools of the Mind", V. Stibic, 1982.

2.  a.  You could derive your own 'controlled vocabulary' for classification
and/or indexing based on your accumulated experience to date and fit
everything to it.  If you are an amateur at the task, you would probably
make some of the mistakes that an expert would avoid.
    b.  You can create files using the most obvious primary classification
words and subdivide or reclassify based on your own experience, as needed.
You could supplement with computerized cross-references.  You would then
have to decide on whether the cross-references should be at the level of
folders or individual documents.

Dennis C. During
dcdur...@gmail.com

I must admit to being slightly confused with your personal taxonomies
issue. DT has excellent (even unparalleled) search, but also storage
in heirarchies as well. Like the filesystem in Mac OS X 10.4, we can
store things in a heirarchical folder structure or do fuzzier searches
as well.

I suppose I see del.icio.us tags as a middle step. It's not as highly
structured as a folder heirarchy, but does provide some structure. It
seems actually less useful in the retrieval-- let's face it, it's a
hassle to page through Merlin's 21 screenfulls of del.icio.us tags.
And alphabetical order tells me nothing about context: how does Amazon
relate to Analog or BBEdit with Begging? Nothing; they're next to each
other simply because of how they're spelled.

Where tags shine is in the creation: I need to remember no paths nor
precedents. If it's about Shakespeare, tag it "Shakespeare." Sure,
maybe you already have "Shakespearean" and "Theater:Shakespeare," but
none of us want to be hassled to consider that fact while we're
creating a document.

Which is, I think, the real beauty of an inbox. Create a file (or a
snippet, or whatnot) and toss it into the inbox. Don't worry about
filing it at all. Later, in a processing step, you can thoughtfully
decide to put the file in the correct location(s). Voila: a taxonomy
which accurately reflects your vision, the context of the item, and
how you're likely to want to retrieve it.

Or am I missing something?

If there's one flaw in DT, I think it's breadth. It does Wiki, but not
nearly as well as VoodooPad. It does outlining, but not nearly as well
as... OK, it is amazingly sucky as an outliner. What's the killer
feature of DT? For me, it's the "See Also..." fuzzy logic stuff. It
helps me see relationships which I otherwise might not see. No, not
every day nor with every document, but it will occasionally make me
say, "Hey.... that's a good point," as if the computer were actually
collaborating with me with actual human intelligence. Kinda spooky, in
fact.

That's the top-drawer, justify-the-expense reason, but the more common
use case for me is as an integrated filesystem and lightweight word
processor. Yes, I have a FS and a word processor (a few of them,
actually) on my computer already, but DT combines them so elegantly
and pleasantly, it makes me happy (and productive) to use it instead.

DEVONthink is a funny app, I think. It's not a killer app; almost
everything it does is replicated in other applications. It is not
simple (else they'd have decent marketing to describe what it does)
nor is its value obvious. But it has wormed its way into my workflow
to make itself the only permanent icon on my Dock. When I find
something I want to save for later, I CMD+SHIFT+) without even
thinking about it.

But back to Emory's other question: does anyone actually use
DEVONagent? What for?

That said, I caught myself downloading it again yesterday. I played
around with it for a half hour, and then realized I just didn't have
the time to figure it out.

This way, if I'm trying to find out, say information on mortality
rates from salmonella, I can pull in a huge load of pages, most of
which have noting to do with the subject I want (much like your
average google reply to a vague search), and then I can use DA to comb
through the results using a number of filters I can tweak as I go, and
show me intelligent summaries of the "relevant" articles.  This
usually results in a pretty good speedup in these sorts of big
searches.  And of course, it's a breeze to capture the resulting pages
to DT for later reading and analysis.

Of course, it's not perfect.  Often it's filters perform too loosely,
finding, for example, search terms in unrelated parts of the
documents, and so on.  And I've asked a couple of times for each of
the different filter terms to be hi-lighted with their own color in
the summaries, to give better visual feedback for quick scans of the
results.  And despite being promised this seemingly small feature back
in DA 1.1 and then 1.2, we are now on version 1.7, and we still don't
have it...  I guess I need to bug Christian again.

It depends on the size of your system, obviously, but I think if you
can roll your own top-level vocabulary and keep that pretty much
intrenched in stone, then you can afford to play fast-and-loose with
the lower levels of the heirarchy to some extent.  Being sure about
that first level minimises the amount of searching you will have to
do.  And, in the final analysis, successfully rolling your own gives
the best possible outcome.  (For personal filing and categorising, I
mean, not for the Library of Congress.)

Andy.
--
Give me ambiguity, or give me something else.

So...anybody got any suggestions?
thanks
Ben

  That's a neat idea actually.  If there's no program available for
that I'd try keeping a folder with clippings in Documents, or even
having the clippings available in Stickies.  I think I'll try the
folder with clippings route since I'm going to try that as well now
that you've mentioned it :)

--
simon.c...@gmail.com