公開
DjVu, CBZ, and the Long Tail of Document Formats
Two formats you'll actually run into: DjVu, built for compressing scanned pages, and CBZ, a renamed ZIP file for comics. Neither is obscure once you know what's inside.
What it is (one-paragraph answer)
DjVu and CBZ solve two unrelated problems. DjVu compresses scanned pages — mostly text — down to a fraction of what a plain image would take. Built at AT&T Labs between 1996 and 2001, it splits a page into a low-resolution color layer and a separate high-resolution layer for the text itself, each compressed with a codec suited to that kind of data. The payoff: a couple hundred scanned pages, sharp enough to read comfortably, can fit in under a megabyte. CBZ is much simpler, and not really a format at all. It's a ZIP archive with the extension changed, containing a folder of page images in order. Comic book archives work the same way whether the extension reads .cbz (ZIP), .cbr (RAR), .cbt (TAR), or .cb7 (7z) — unzip one and you'll find plain JPEG or PNG files, one per page. The format lives in the reader, not the file.
Why it matters
Neither of these is a museum piece. DjVu was the format the Internet Archive used for its Million Book Project starting in 2002, so a real slice of what got digitized from libraries in that era — theses, out-of-print monographs, older scans nobody has re-processed since — still exists only as DjVu, not PDF. Gone looking for material worth reading from the Internet Archive and downloaded something that wouldn't open the way you expected? This is often why. The trade-off is worth knowing up front: DjVu is fixed-layout, so a page displays at one fixed geometry no matter your screen size, and its text layer uses a compression trick that can occasionally swap a similar-looking character for another — a 6 read as an 8, say — with nothing in the file flagging that it happened. Comparison research on e-book formats says plainly that DjVu files cannot reflow the way EPUB text can. It's the right tool for preserving a scan faithfully. Adapting text to a small screen isn't what it's for.
Where you'll see them
DjVu turns up in academic archives, older library scan collections, and — even though the Internet Archive stopped accepting new DjVu uploads in February 2016 — plenty of material it digitized years earlier that was never converted. Wikimedia Commons still conditionally accepts DjVu uploads alongside PDF. CBZ and CBR live in a different corner: digital comics and manga, especially collections people assembled or backed up themselves rather than bought through a single publisher's app. Moved a comics library off a storefront that shut down? There's a good chance what you're left holding is a folder of CBZ files. Both formats sit near the fixed-layout files that don't resize — they preserve a page exactly as scanned or drawn, the opposite trade-off from what EPUB 3 is built to do.
How Margina handles it
The point isn't to convert a DjVu scan or a CBZ archive into something else before you can read it. It's to open the file as it already exists, however you came to have it. Margina reads what you send yourself, without asking for an account or routing the file through a server first. Same idea behind how Margina treats Kindle's own closed formats: meet the file where it is, rather than making conversion a precondition for reading. An old scanned thesis, a chunk of an archive's older backlist, a comics folder inherited from a defunct app — that's exactly the kind of long-tail file this is meant for. Margina is available on the App Store and Google Play.
Frequently Asked Questions
Is DjVu the same thing as PDF?
No. Both can hold scanned pages, but DjVu splits a page into a separate text-mask layer and a background/color layer, each compressed with a codec suited to that kind of data — which is why a DjVu scan is often far smaller than an equivalent PDF built from plain page images.
Can I open a CBZ file without special comic reader software?
Yes. A CBZ is a ZIP archive with a different extension, so any unzip tool opens it and reveals a folder of plain image files. A dedicated reader just adds page-turning, zoom, and library organization on top.
Why would I still run into DjVu files today?
Older scanned books, academic papers, and library archives — including a large slice of what the Internet Archive digitized in the 2000s — were saved as DjVu before PDF became the universal default, so the format still turns up in research and archival contexts.
Does Margina convert DjVu or CBZ files before I can read them?
No conversion step is required — you open the file as it is. Margina reads the format directly rather than asking you to convert it to something else first.
Conclusion
DjVu and CBZ aren't edge cases so much as the actual shape of what people have accumulated over twenty years of scanning, downloading, and archiving — a thesis saved before PDF took over, a comics folder carried from one defunct app to the next. A reading app's job is to meet you at the format you already have. Not to make you convert first.