Slashdot Mirror
Are Digital "Margin Notes" Possible Yet?
Stavo asks: "I'm looking for a robust, reliable personal knowledge management solution. As a professional researcher, I read a lot of text-based content. I prefer to mark up content, by underlining or adding margin notes. I also need to retrieve and search content. The low tech solution is printing the text and using a pen to mark up, then filing the papers. If I want to quote a source, I have to type the quote. With the advent of Tablet PCs and similar tech, I'd like to find a way to keep the content digital. In other words, if I download an journal article in PDF or HTML, how can I mark it up, save it, and later search/retrieve it? Shouldn't computers provide a better solution than voluminous file cabinets filled with dead trees?"
Unless I'm missing something, the full version of Adobe Acrobat can do all that. Annotations in text, voice, file attachments, etc. and a file indexing service "Adobe Catalog". Any PostScript output can be turned into a PDF, there are even free tools to do this on Linux. But if you're using Macintosh or Windows, you can print directly to PDF format. Acrobat 5 can even render web pages into PDF format, preserving links. IIRC Adobe also has a fully functional time limited demo available.
Now, getting those dead-tree file cabinets into PDF format is another problem alltogether. Possibly using overseas data-entry companies?
Yep, head on over to www.adobe.com and research.
From what it sounds like, though, he'd have to print all documents to distiller (in winblowz) to use that in all the files he'll be working with.
"There is a way that seems right to a man, but its end is the way of death." Proverbs 16:25 (NKJV)
It's an app used primarily in the legal industry. You can in hard copies or import text/doc files. Once the file's been imported into the system, you can highlight bits of text and do the things that you need to do. Used it a lot when I was a legal assistant, mostly for summarizing deposition or trial transcripts.
Check 'em out here at http://www.summation.com
-- anthony
It's definitely possible, as others have said, Adobe Acrobat already does this. I happen to own a copy of Acrobat, so I've had the opportunity to play around it's capabilities.
Short answer -- it works pretty damn well. But not with a mouse. A mouse just isn't suited to making marginal notes (i.e., checking an important idea, underlining a particular phrase, or circling an important passage). A tablet device with a stylus, however - that holds promise.
Other things to note: Acrobat provides two types of commenting systems. First, notations -- you can hilight, underline, circle, or freestyle directly onto the document. Second, "sticky-note" style comments. One very cool thing about the sticky-notes are that they render translucent so that you can still read the text underneath the note.
Also, as far as I can tell, the commenting systems appear to be embedded into the document as PDF code. Specifically, gv is able to render notations (hilighting, underlines, etc). gv is not able to render the sticky-notes, however. I don't know if that's because gv simply can't handle the sticky-notes or because the sticky-notes are in some type of proprietary format. xpdf doesn't render either form of comments.
So, if you're using Windows, are comfortable with proprietary software, and can afford $250, you're more or less set (assuming that pen computing lives up to its promise).
Things get a bit more tricky if you're looking for free-software solutions. As far as I know, there's nothing out there as of yet. And I don't know how difficult it would be to implement (I do know that it's way beyond my capabilities, however). But because it appears that Acrobat embeds the comments as native PDF code, it should be possible. The question is whether or not anyone's willing to take up the cause...
Acrobat Reader can.
gv, ggv, and gsview cannot.
Come to think of it, the Open Source world has seriously missed the ball in general when it comes to PDF documents. Open Source PDF viewers suck. In every single Open Source PDF viewer I've used, I've run into documents where the renderer has the orientation wrong -- and not just the orientation, but the "orientation of the bounding box" being different different from the "orientation of the drawn data on the bounding box", so that the top and bottom of the drawn data is lopped off, and there's a ton of white space to the left and right.
The only Open Source PDF viewer I've used that can (gasp) search for text is gsview, and it's *really* flaky and doesn't highlight found text. Nothing like trying to read through a page of text to find the one word you're looking for.
I've never used an Open Source PDF viewer that can antialias embedded bitmap images, which makes things look awful and unreadable.
Finally, Acrobat Reader for Linux is completely awful, and leaks memory like a sieve. I have a friend with about a gig of RAM that Acrobat Reader sucked through in about six minutes of dragging and scrolling the document.
Since PDF-viewing is one of the major office activities (along with world processing, and email), this is an enormous impediment to the use of Linux (or any UNIX) in a desktop environment.
It's extremely embarrassing to say something nice about Linux, have a friend use it, and then realize how truly much Linux software sucks at handling PDFs. "You mean I have to read through this thing manually instead of searching?" "Why does this print turned sideways? It works fine on Windows!" "Why does this look so bad?"
I predict Linux will not take off on the office desktop until (a) OpenOffice doesn't look and work completely differently from every app out there, and is free of cosmetic bugs *and* handles MS Office documents almost flawlessly, and (b) PDF viewing doesn't suck.
And for home use, (c) until the Linux sound architecture doesn't completely suck. Right now, the only way to obtain software mixing is through a dropout-prone, non-real-time-scheduled sound server with lousy latency. They usually don't share the sound device very nicely, either. Many sound systems can't do hardware mixing. Linux doesn't have a single way to do software mixing fallback, where a user out of hardware channels will automatically do real-time-scheduled software mixing. Pretty lame. Oh, and at least esd has truly awful resampling. Usually, when new users come to Linux, I hear "why is my sound dropping out when it doesn't on Windows", "why is there lag between something happening and a sound playing", "why does my sound sound so bad (this when resampling is occurring", or "why can't I hear ICQ sounds when xmms is playing?"
May we never see th
CMU's AUIS (or whatever they took to calling the expanded Andrew Tool Kit stuff) included a tool very similar to this, ~10 years ago. Students could use a word-processor-like program to write papers (or import them from other sources), then submit them on-line to a central server. Teachers/TAs/whatnot could read, edit, and markup (that is, either change the document or annotate it) on-line, and share the results with other teacher/TA/whatnots and/or the students. It was a pretty useful idea, used for a couple MIT classes, but the implementation was pretty flawed (the server crashed often, and was prone to losing files), so they stopped.
Come to think of it, that's pretty much the experience with AUIS/ATK in general. In significantly less-nice terms, a friend of mine once said:
like all CMU code: way cool design, implementation like wet camel shit.
At this point, I believe that AUIS is pretty much defunct, so I doubt anyone cares. The code is probably available under OSS license if anyone cares (I believe it was old-style BSD (with attribution)).
Perhaps I am simply luckier, but I have never had xpdf get the bounding box wrong in either fashion you describe.
Regardless of luck, I search for text in xpdf without trouble* at least several times a week, for months.
Check it out, at freshmeat, for example.
* by `without trouble', I don't count xpdf's nearly overbearing ugliness as `trouble'. :-)
For articles that you scan from the printed world and store as scanned images, I would recommend something like PaperPort Deluxe. In addition to offering nice folder-based scanned image management and editing, it also allows annotating these scans via virtual sticky notes, text boxes, free-form drawing and highlighting, etc. All these annotations are stored in a separate layer but can be permanently "burned" into the underlying raster image at any time. In addition it also offers background full-text indexing (after OCR-ing on the fly) and searching. It's quite space efficient with scans, especially when scanning at 300 dpi lineart (which is most useful for archiving printed articles, since they can be printed again at decent quality), with the average magazine article page taking up only about 30-40KB as a compressed TIFF image.
Anyway, while it also lets you manage Word, Excel, PDF etc. files and web pages (and view them within its interface), unfortunately it won't let you annotate those. That would indeed be a very nice extra feature, maybe it should be suggested to ScanSoft. But still, scans of printed articles do make up a very substantial subset of research articles (my wife also does research and has to deal with this same issue), so PaperPort's features are still very useful. Plus it's a very inexpensive product, often included for free with $40 scanners.