Slashdot Mirror


User: dmarney

dmarney's activity in the archive.

Stories
0
Comments
3
First seen
Last seen
Profile
(view on slashdot.org)

Comments · 3

  1. DjVu Does What PDF Can't on A New Web Image Format · · Score: 1

    Here are just a few of the things that DjVu can do that PDF cannot:

    1. Encode a 20MB color scan into a 100-200K file, and still produce a good-looking image.
    2. Encode a 300 dpi B&W scan into 17K, and still be able to read the image.
    3. Create separate image and data layers.
    4. Create a multi-page document either as a single file, or as a linked list of indivdual pages, with a file for each page.
    5. Construct a URL that will drop a user into the middle of a document (page 50 of 100, for example), without losing the context of the other pages.
    6. Provide automatic installation of the free DjVu viewer without sign-up forms that harvest my customer's email addresses
    7. Zoom and pan on the fly.
    8. Convert to other file formats such as JPEG.

    For my money, DjVu is a superior file format that has been nicely tuned for the web. PDF was optimized for print pre-press, and it shows.

  2. Re:Some facts off the top of my head... on A New Web Image Format · · Score: 3

    I have been using DjVu for more than a year now, and have tested it extensively against other image compression technologies. If you've got scanned images to display on the web, especially if they are in color, then DjVu is far and away the best technology on the market. Here is just a short list of the things that I like about DjVu: 1. I can encode a 20MB color scan into a 100-200K file, and still get a great-looking image on the screen. I can create astonishingly tiny B&W images (how about a full page of text scanned at 300 DPI rendered in 17K?) Nothing else comes close. 2. I can create separate image and data layers. DjVu produces a color background layer, color foreground layer, high-contrast B&W layer, and a data layer. This is essential for doing OCR, where we really need to have that B&W image to get to the text. 3. I can encode the DjVu image to automatically upsample to match the greater resolution of my printer vs. the screen, in the same file. 4. I can create a multi-page document either as a single file, or as a linked list of indivdual pages, with a file for each page. No more horrid PDF byte-serving! (Please pardon us as the Author weeps for joy.) 5. I can construct a URL that will drop a user into the middle of a document (page 50 of 100, for example), and not lose the context of the other pages. 6. I can use the EMBED tag to provide automatic installation of the free DjVu viewer, and I get to specify which image comes up once the software is installed. There are no sign-up forms, no harvesting of my customer's email addresses, and no taking the user out of my visual space. 7. The viewer zooms and pans on the fly. You can zoom in to 100%, and pan through the image simply by clicking and dragging the mouse. This is the only way this should be done. 8. All of the encoding and decoding tools are completely free for eval purposes. The decoder is free for all purposes, and most of the source is open. 9. Once I create a DjVu file, I can convert it to other file formats such as JPEG. Try doing that with PDF! 10. IT LOOKS BETTER -- A LOT BETTER -- THAN ANYTHING ELSE.

  3. Re:The Ability to Search on A New Web Image Format · · Score: 1

    DjVu has both image layers and a data layer. The high-contrast b&w layer is OCR'd and the recognized text is stored in the data layer. The image can thus be indexed by any search engine.