Welcome to axw3, a http proxy and AX25 gateway for Linux that allows your browser to fetch web pages from the Internet using a pure AX25 connection. IP encapsulation is avoided, therefore IP and TCP headers, name queries and TCP retransmission are unnecessary, so you'll save some bandwidth. It was tested only with the tools from the Linux ax25-utils package (call and node), but it will probably work with others, for instance when the remote node is running TNOS or JNOS. The diagram shows its operation:
WWW Browser | /|\ I want | | www.y.com/path | | filtered page | | \|/ | axw3 | /|\ exec call | | send commands | | node msgs + call warnings + page | | \|/ | call | /|\ connect node | | forward commands | | node msgs + page | | \|/ | node | /|\ telnet www.y.com 80 | | GET /path | | page | | \|/ | www.y.com
The very nature of this method is not appropriate to download binary data (images). In fact, if you try make that, both axw3 and call may die. For the moment we suggest disallow images autoloading in the browser.
Installation instructions are included in the distribution file. This code is under the terms of the GNU GPL license. You can download axw3 from ftp://ftp.ime.usp.br/pub/ueda/axw3-0.2.tar.gz . Send questions and comments to email@example.com (Ricardo Ueda Karpischek).
Ricardo Ueda Karpischek wrote axw3 from scratch. APPR (www.appr.org.br) provided the necessary background and expertise to develop this software.
1. only http:// URLs are accepted.
2. only GET requests are accept (so axw3 currently cannot handle FORMS).
3. axw3 may not recover from some errors. In these cases you'll need to restart it (maybe necessary kill manually the call process).
4. A bunch of lines generated by the browser query are ignored, in particular those regarding the date of cached pages.
5. The stop button of the browser usually will not work as you expect.
6. Download of images may provoke call and axw3 death.