This is a README for the font compression reference code. There are several
compression related modules in this repository.

brotli/ contains reference code for the Brotli byte-level compression
algorithm. Note that it is licensed under an Apache 2 license.

src/ contains prototype Java code for compressing fonts.

cpp/ contains prototype C++ code for decompressing fonts.

docs/ contains documents describing the proposed compression format.

= How to run the compression test tool =

This document documents how to run the compression reference code. At this
writing, the code, while it is intended to produce a bytestream that can be
reconstructed into a working font, the reference decompression code is not
done, and the exact format of that bytestream is subject to change.

== Building the tool ==

On a standard Unix-style environment, it should be as simple as running “ant”.

The tool depends on sfntly for much of the font work. The lib/ directory
contains a snapshot jar. If you want to use the latest sfntly sources, then cd
to the java subdirectory, run “ant”, then copy these files dist/lib/sfntly.jar
dist/tools/conversion/eot/eotconverter.jar and
dist.tools/conversion/woff/woffconverter.jar to $(thisproject)/lib:

dist/lib/sfntly.jar dist/tools/conversion/eot/eotconverter.jar
dist.tools/conversion/woff/woffconverter.jar

There’s also a dependency on guava (see references below).

The dependencies are subject to their own licenses.

== Setting up the test ==

A run of the tool evaluates a “base” configuration plus one or more test
configurations, for each font. It measures the file size of the test as a ratio
over the base file size, then graphs the value of that ratio sorted across all
files given on the command line.

The test parameters are set by command line options (an improvement from the
last snapshot). The base is set by the -b command line option, and the
additional tests are specified by repeated -x command line options (see below).

Each test is specified by a string description. It is a colon-separated list of
stages. The final stage is entropy compression and can be one of “gzip”,
“lzma”, “bzip2”, “woff”, “eot” (with actual wire-format MTX compression), or
“uncomp” (for raw, uncompressed TTF’s). Also, the new wire-format draft
WOFF2 spec is available as "woff2", and takes an entropy coding as an
optional argument, as in "woff2/gzip" or "woff2/lzma".

Other stages may optionally include subparameters (following a slash, and
comma-separated). The stages are:

glyf: performs glyf-table preprocessing based on MTX. There are subparameters:
1. cbbox (composite bounding box). When specified, the bounding box for
composite glyphs is included, otherwise stripped 2. sbbox (simple bounding
box). When specified, the bounding box for simple glyphs is included 3. code:
the bytecode is separated out into a separate stream 4. triplet: triplet coding
(as in MTX) is used 5. push: push sequences are separated; if unset, pushes are
kept inline in the bytecode 6. reslice: components of the glyf table are
separated into individual streams, taking the MTX idea of separating the
bytecodes further.

hmtx: strips lsb’s from the hmtx table. Based on the idea that lsb’s can be
reconstructed from bbox.

hdmx: performs the delta coding on hdmx, essentially the same as MTX.

cmap: compresses cmap table: wire format representation is inverse of cmap
table plus exceptions (one glyph encoded by multiple character codes).

kern: compresses kern table (not robust, intended just for rough testing).

strip: the subparameters are a list of tables to be stripped entirely
(comma-separated).

The string roughly corresponding to MTX is:

glyf/cbbox,code,triplet,push,hop:hdmx:gzip

Meaning: glyph encoding is used, with simple glyph bboxes stripped (but
composite glyph bboxes included), triplet coding, push sequences, and hop
codes. The hdmx table is compressed. And finally, gzip is used as the entropy
coder.

This differs from MTX in a number of small ways: LZCOMP is not exactly the same
as gzip. MTX uses three separate compression streams (the base font including
triplet-coded glyph data), the bytecodes, and the push sequences, while this
test uses a single stream. MTX also compresses the CVT table (an upper bound on
the impact of this can be estimated by testing strip/cvt)

Lastly, as a point of methodology, the code by default strips the “dsig” table,
which would be invalidated by any non-bit-identical change to the font data. If
it is desired to keep this table, add the “keepdsig” stage.

The string representing the currently most aggressive optimization level is:

glyf/triplet,code,push,reslice:hdmx:hmtx:cmap:kern:lzma

In addition to the MTX one above, it strips the bboxes from composite glyphs,
reslices the glyf table, compresses the htmx, cmap, and kern tables, and uses
lzma as the entropy coding.

The string corresponding to the current WOFF Ultra Condensed draft spec
document is:

glyf/cbbox,triplet,code,reslice:woff2/lzma

The current C++ codebase can roundtrip compressed files as long as no per-table
entropy coding is specified, as below (this will be fixed soon).

glyf/cbbox,triplet,code,reslice:woff2


== Running the tool ==

java -jar build/jar/compression.jar *.ttf > chart.html

The tool takes a list of OpenType fonts on the commandline, and generates an
HTML chart, which it simply outputs to stdout. This chart uses the Google Chart
API for plotting.

Options:

-b <desc>

Sets the baseline experiment description.

[ -x <desc> ]...

Sets an experiment description. Can be used multiple times.

-o

Outputs the actual compressed file, substituting ".wof2" for ".ttf" in
the input file name. Only useful when a single -x parameter is specified.

= Decompressing the fonts =

See the cpp/ directory (including cpp/README) for the C++ implementation of
decompression. This code is based on OTS, and successfully roundtrips the
basic compression as described in the draft spec.

= References =

sfntly: http://code.google.com/p/sfntly/ Guava:
http://code.google.com/p/guava-libraries/ MTX:
http://www.w3.org/Submission/MTX/

Also please refer to documents (currently Google Docs):

WOFF Ultra Condensed file format: proposals and discussion of wire format
issues (PDF is in docs/ directory)

WIFF Ultra Condensed: more discussion of results and compression techniques.
This tool was used to prepare the data in that document.