search for single repository

Posted by jerryk
Oct 31, 2008 at 04:16 PM

Given that ultrarecall (UR) is no longer being developed, it seems timely to ask the perennial question of which program best allows us to manage the surfeit of information associated with our projects, reference materials, and research (what I call knowledge). At a minimum, that program must be able to assimilate: web page, emails, files, snippets of text. Some meta-data is useful, but for most folks, a highly structured database seems not to be necessary, and that is the case with me as long as I have full text-search.

I have tried to implement a single repository notion in the past with Zoot, Content Saver (now Web Resesarch), Ultra Recall, OneNote. I have substantial experience with each of these programs. Over the summer I had decided to leave Ultra Recall and go into One Note, but just recently decided to test out Web Research again, and am thinking that might be a workable solution. I want to share some of my thoughts, so that others can point out where they differ and point me to better software or processes.

*** Web ****
Zoot: does not yet capture rich text. There is an archive to mht function, but the last time I checked it was buggy. In any event, no editing (highlighting) of the web page permitted, which is a substantial negative.

UR: I know development has stopped on this, but the program still functions. Also the comparison may be illuminating. UR was always terribly slow in capturing long web pages. Editing was added, but the UI was always a bit cumbersome. Also search hits were not highlighted.

OneNote: Saving clips of a web page was fine, but sending entire pages produced bad copies. One way around this was to grab screens using SnagIt, but that often created bloated image files (which were OCR’d by OneNote).

Web Research: For me, the fastest web capture, and easiest filing (besides potentially Zoot’s auto-filing folder rules). Sometimes forms, etc., don’t get captured correctly—but for those sometimes only a screengrab will do.

I think WR wins this one.

*** Emails ***
Given full text indexing of emails, I do not file and instead pile. I used to use NEO b/c of its convenient correspondent folders, but I’m trying to cut down the # of apps I’m using. So I’m experimenting with just using outlook and Vista’s Windows Desktop Search (WDS).

Zoot: Great at sucking in emails. At one time, I used to use smart folders to automatically file emails, etc. But again, with good enough searching, I’ve decided that I generally don’t have to file emails. The most important ones, I can “file” or save in some project folder separtely.

UR: Fine at storing emails (although I never liked the fact that attachments were never shown separately from the *.msg file).

OneNote: printing out emails took too many pages, and did not allow for quick and easy navigation. Emedding the email was another possibility, but OneNote has problems with embedded items. First, WDS doesn’t search embedded files. Second, the icon is large and doesn’t show the entire name.

Web Research: With the add-in, saves emails fine. It would be nice if somehow attachments appeared as subdocuments of the email—the way that UR allowed anything to be a parent/child of anything else. But it seems to work fine enough.

*** Files ***
Zoot: allows file links, not storage of files. This allows annotation with keywords, but if you want full text indexing, you have to run a WDS (or comparable) search as well. This seems like 2 repositories—not a great thing. That said, there’s a file folder sync that allows any new files stored in an OS folder to create a new zoot item, which can be tagged.

UR: storage of long files was slow. Make a minor change, and it would reindex. Some problems with data loss if the file you were working on didn’t close down right or if UR closed down first.

OneNote: Easy to work with embedded files. Changes are saved within the onenote page. But as explained, no WDS search of the embedded files. Also the UI of moving around embedded file icons was clumsy. This is an example where the metaphor of a network page ends up breaking down.

Webresearch: Can “store” files into its DB. But large files (say 100mb) seems to choke it. Also, in order to edit the file, you have to export it to a OS folder, save the file, then reimport. The exporting is clumsy (can’t click and drag). That said, the files are indexed by WDS.

If files could be edited like OneNote but appeared in UI like WR, I’d be happy.

*** clips of text (e.g., passwords, #s) ***

Zoot: phenomenally good at this

UR: fine

OneNote: fine

WR: can create a new “note” to type this stuff in

——————————————————-

With all this, here’s how and why I ended up with Web Research tentatively.

UR is out because it’s no longer developed (also Web storage was slow, and frankly the UI was always more difficult than necessary).

Zoot, which I admire greatly, still can’t deal with rich text. Using various linking technologies adds complexities, changes that links can get broken with moving computers and files, and the like. (I’m using Zoot for a more narrow set of tasks only.)

OneNote. I thought this could work. Even if embedded files could not be searched, I presumed that the next version would add that functionality. And by embedding files as subitems in an outline, I could collapse the outline and manange complexity that way. But the paper notebook analogy fails once you are managing 20+ files for a project. Printout out entire pdfs, etc., into OneNote could be hundreds of MBs, so it must be the embedded file strategy (I use Pdf annotator to mark up files.)

Web research. Best at web. Competent with emails. Competent with text snippets. Weak with files (can’t edit stored files without exporting/reimporting). That said, WDS will find everything on your hard drive and within a WR database. One can create hyperlinks to a single WR item (which I can use with Mindmanager, to track support material for projects.) And the UI is the best of the lot.

My return to WR has only been about a week, and I’m groaning at the transition costs of moving data over. But OneNote isn’t working for me. What are your thoughts and reactions?