A slight amount of waste is intrinsic to HTML, because HTML is written in ASCII text. ASCII is defined as using only seven bits of each byte, so one bit in eight, or 12.5%, is wasted. A larger amount of waste is due to the fact that text is highly compressible, but no compression is used for most HTTP transfers. It is normal for a text compression program to reduce the size of a text file by half, meaning that file can then be downloaded in half the time. Right now, transmission bandwidth is the bottleneck and CPU power for decompression is cheap, so compressing web pages would seem to make sense, even if it would make debugging problems harder.
Lack of compression is not an issue for small pages, but large text
pages can be sent significantly more quickly in compressed format to
a browser that understands
gzip compression, for
example. (I believe Mosaic is the only popular browser that
gzip.) Another option is to compress
your content and configure your web server to use a certain MIME type
for that compressed content, but you then have to ask your users to
configure their browser to launch the decompression utility when a
file with a certain content type is received. This requires a bit of
work on both the client and server side.
When composing HTML, try to keep pathnames short, both in number of directories in the path and in the length of each directory name.
You can scale static ...