File tagging/searching (Was: Now, if DOpus did THIS...)

jon wrote:
The integrated Google Desktop Search in Opus will only return real files - it doesn't look in email messages, saved webpages, etc.[/quote]

Thanks Jon.

That certainly explains it. There was only one real file in the list.

Heads up & FYI -- I am pasting what I commented to Google upon uninstalling Google Desktop 2:

"Well, what can I say...

After using the program for a while, I discovered that all my previous searches may have been incomplete.

You see, today I wanted to open some documents from my My Documents folder. Guess what? Nothing was found. I had at least 10 documents with similar filenames, but they were not coming up in GDS, even as I was looking at them listed in Explorer.

So, after some research, I discover to my AMAZEMENT that GDS only indexes 100000 files in each drive!!!

To say this is RIDICULOUS would be an understatement. In this day and age of 300GB hard drives, the king of completion in search gives the users of its desktop product incomplete (or zero, as in my case) results. Great.

You can only imagine my level of confidence in GDS now (and the results I get from it).

Whoever made this bone-headed marketing decision (because I am sure that's what it is, there's no technical reason for this) should be fired. Period.

I used to be a true fan and supporter of Google, but this really shook me. And you know why? Because it's misleading--you guys should make this limitation LOUD and CLEAR--people shound KNOW about this. I am sure many people get a FALSE SENSE OF CONFIDENCE using this very broken program.

Like I said, I will never know how many of my previous searches were incomplete--and what those searches were about and what the implications might be that I didn't find the information I was looking for at the time I needed it.

I never expected such low standards from Google. Really."

On the other hand, ignorance can be bliss :slight_smile:



A quick web search (in Google, no less) turned this related blog entry, by none other than Mark Russinovich of Sysinternals fame: ... sktop.html

Mark says

"What good is a partial search of my documents, especially when I have no control over the portion that's searched? Does Google's web search engine only index parts of web pages?

Its a little amazing to me that, besides not clearly telling you what it indexes, Google Desktop doesn't have options that allow you to specify the following options for the indexer:"


"I don't ever see myself interested in a partial documents indexing".

I couldn't agree more.

I've been following this thread since it began, and I must say I'm quite intrigued. I'm starting to think that this is very interesting stuff.

I'd be very interested in seeing how you use Evernote. I've downloaded it based on the reccomendations in this thread, and I'm liking it so far.

Also, I'm a bit confused on one issue. Why are flat tags (as opposed to hierarchical tags) insufficient? From how I'm understanding things (as an admitted new-comer to classification systems), as long as an item was fully and properly tagged, intersections would render hierarchy useless. Of course, if you took your example paper and tagged it only as "Aristotle" and not under "philosophy," then intersections would do you no good. Proper tagging would be key, but as long as there was no limit to the number of tags that an object could have...

Perhaps an alternative to hierarchies would be a system where anything tagged with "Aristotle" would automatically also be tagged with "philosophy" and "Greece," being that Aristotle was a greek philosopher. Using intersections for tag creation as well as for browsing would eliminate the need for hierarchy, would it not?


chriscrutch, I am putting together a screencast of my use of EverNote. You will see soon how I envision DOpus to work, based on my use of EverNote.

As for flat vs. hierarchical tags, I believe that hierarchy gives more flexibility and another dimension to tags. I am certainly not a disliker of hierarchies (although I realize I may have made it sound that way). What I do dislike is not having a choice for the right tool for the job. Hierarchies certainly have their place and usefulness, and with hierarchical tags I feel that we get the best of both worlds. Again, you will see this in my forthcoming screencast of EverNote.

Just to give you two quick examples, one "right" and one "wrong":


This hierarchy example, in my classification cosmotheory, is wrong. Why? Because "Greek" is a much more general tag than this implies. This hierarchy implies that for something to be "Greek", it has to be about "Philosophy", which is clearly wrong. "Greek" is a much more broad concept that involves (and has to do with) a lot more than Philosophy.

Another example:

--Visual Basic

Now, I feel this is about right. For something to be about "C++", it has to be about programming. It cannot be about something else. This way, I could choose to browse about ALL programming related stuff, not just C++. I couldn't do that with non-hierarchical tags (unless I also tagged everything with "Programming" as well, which cognitively is a lot of work, and I will, more likely than not, forget to do it).

Of course, I would also tag it as something else, for example:

What is it?
--Sample code

So, yes, something is about "C++", but what exactly is it? It could be actual sample code, it could be a tutorial, or it could be an article about the usage of "C++" in the marketplace.

So, in a nutshell, the more tags you give to an item, the more "dimensions" you introduce to it, and to the ways you can browse for it. So, I believe that hierarchical tags eliminate the need for giving a whole lot of tags (which is messy and a lot of work), and help organize many concepts together. This leads also to good, logical organization of tags that make logical sense. It cleans up, tidies and organizes your tag landscape, so to speak.

Last but not least (and this may be the most important point here), without hierarchical tags, you CAN'T HAVE faceted classification!

More later.



I use Photoshop album which has a tagging system to tag images and later retrieve them. That has become one of my favorite programs to keep track of my 10000+ images.

Another interesting hierarchical approach to keeping track of "stuff" such as URLs, files, applications, documents, images, etc., is a thing called BookMark Buddy. Despite it's rather naff name I found it to be instantly attractive, much the same way I felt the first time I clapped eyes on DO. I now use it all the time as a complement to DO. It may be one of those apps that you either take to instantly or never quite get. I can only speak for my own reaction.

Regards, AB

I think you may take a look at image management programs (I personally use iMatch). You have all possible approaches to your problem there.


I think the general idea behind iLiAS request is based on the fact that while there are certainly programs out there that allow management of data collections based on meta-data, tags, etc... they are data type specific. His position is that this capability is a valuable feature regardless of data type and is should sensibly be filled by a file manager. I tend to agree... the need to manage data in an expiditious manner is why any of us wanted a replacement for Explorer in the first place. It shouldn't matter that we might want to manage music files based on a different specific criteria than images, etc etc. All I want a media player to do nowadays is pipe sound through my speakers! All I want an image application to do is let me either look at or edit a picture! Database? Phfaghhhh... blah.


I've just stumbled onto this old thread after searching for an answer to this dream, which I share with you. Thanks for starting the discussion and explaining things so clearly.

I really [b]NEED[/b] this feature too! I was wondering if you have found any interim solution?

Also, what happened to the screenshot of Evernote that you were going to post?

Kudos (also from the MyLife Organized forum)

Permutations and combinations are not the same thing mathematically.
It would be combinations as the order of the tags is not relevevent unless there is weighting to tag order.
And there are considerably more combinations than 100.
[C=10!/10! + C=10!/9!1! +C=10!/8!2! + C=10!/7!3! +....C=10!/3!7! + C=10!/2!8! + C=10!/1!9! ] = Total Combinations.
I think that's right, it's along time since I did probabilty and stats.

There are 120 combination of just three words out of a total of ten words and another 210 with 4 out of 10......

I'm compelled to post: this idea has been knocking around for years it seems- where is it? Why hasn't it manifested? iLiAS's proposal is not only sound, but desperately needed for proper information management. My whole reason for trialling this software in the first place was to see if it supported metadata-based information rediscovery.


Why hasn't it appeared? Because a bunch of other things were done instead, of course. See the Opus change logs for the thousands of other changes which were done since this thread was created.

If you feel GPSoft's priorities are wrong then tell them. Keep in mind that this feature is not trivial to implement (especially if you want it to work with metadata file formats handled by plugins) and so doing it will mean not doing a bunch of other things which other people have requested. That's always the balancing act. All anyone can do is write to GPSoft and express their desires and then see what happens. If a lot of people ask for the same thing then its priority will presumably be increased (unless it's something GPSoft really don't want to do, of course).

PS: What is "metadata-based information rediscovery"?

I didn't mean to imply the developers were slackers- though I am forced to wonder about their priorities. Maybe it's because my needs are very "power-user" orientated, maybe it's because of the circles I frequent, but there is an absolutely massive calling for software that manipulates and searches by metadata.

I thought that was what I was doing by posting to this thread. :wink: It gives my request a bit of context, and ensures that it's not seen as just idiosyncratic or particular to me- a very large set of information managing power users want to manipulate information by metadata.

Check out Ilias's explanation on the first page of this thread, I think that's pretty adequate. In short though, to help you out: meta-data based information organisation isn't just there so that you can search for specific files (indexing services do that, arguably better), it's so that you can search for a type of information and (re)discover relevant files that you'd forgotten/were filed differently, or even see new connections between existing information. Aristotle's works might be filed under Philosophy and Aeschylus' under drama, but searching by metadata tag "greek" would show me all their works, and many more besides.


Hey there... I think a fair number of ppl agree this sort of capability could be quite useful. But either way... what Nudel was getting at was that you should contact GPSoft to let them know your program suggestions... he linked to the page you should do so in his post.

This is a user-run forum... :wink:

Ah-ha! I'll be sure to contact the developers shortly, then. :slight_smile:

For those following the thread I've found that xplorer2, a competitor product, while not offering nearly the functionality of DOpus, does actually provide a shortcut key combo to alter the "comments" field from the explorer pane, which is sufficient for working with the metadata (I think). I'm playing with both as a result, but at least I can jury-rig a temporary solution. Others might want to in the meantime.

I wish people would stop talking in jargon and actually explain what they want to do.

From your last post it sounds like all you want is a way to assign comments to files. Well, Opus lets you do this - and you can configure any hotkey you like for it, not the one chosen for you by a developer.

Is there a RAW command to modify an ntfs comment though, and not just the description?

Don't think so, but a description should work just as well as a comment, I think.

That's what I'm after. It's not quite as sophisticated as Ilias's conception of how one might work with metadata in DOpus, but I think you can hobble together some kind of system with this approach.