one-day sale on hackable wireless router
D. Hugh Redelmeier
hugh-pmF8o41NoarQT0dZR+AlfA at public.gmane.org
Fri Sep 5 15:24:00 UTC 2008
| From: Lennart Sorensen <lsorense-1wCw9BSqJbv44Nm34jS7GywD8/FfD2ys at public.gmane.org>
| On Thu, Sep 04, 2008 at 11:53:47AM -0400, James Knott wrote:
| > The source code is available from that site, so if you're so inclined
| > you can hack it.
|
| Having source code is not the same as having the ability to generate and
| load new firmware. Just look at the tivos.
True. But it is even worse than that. Even if the system will let
you flash firmware that you have built (i.e. has a flash mechanism and
does not require the firmware to be signed with a key that only the
manufacturer has), there are real practical problems that I've
observed:
- the released software is generally incomplete. Some of the stuff
built on top of Linux is usually proprietary.
Sometimes one can replace this with open source, sometimes not.
Sometimes one can grab the binary of the proprietary stuff from
inside a firmware load and put it in your firmware build. A messy
job.
- Often the build system isn't provided and it is work to recreate it.
If you don't exactly recreate it, you may well expose latent bugs in
the source code.
Often the build system requires a particular (unspecified) host
system. That may be awkward to discover and provide.
- Some systems have no recovery system for when you flash buggy
firmare. If you build it, it will be buggy.
- Too often, the released source doesn't actually match the released
firmware.
- the community of interested hackers is too small to spread the
required development load
I think that all of these have happened in the case of the Linksys
WRV200. A great disappointment to me.
I bought one on the strength of the downloadable manual. It had a
described an IPsec feature that was only implemented by FreeS/WAN and
its successors. I wanted this feature and I was amused to buy
something off-the-shelf with code that I wrote.
- the feature was not, in fact, available. Because the GUI didn't
support it (the code underneath certainly did). When I phoned
support, they said the manual was wrong. I should have returned it
there.
- they dragged their feet releasing source. Even when they did, it
was for old versions.
- the system had and has bad reliablility. I could probably have
fixed that (for free, with source code). The Openswan folks
approached them offering consulting help (for money) but were not
used.
- there is a community of complainers. All share voodoo tips to make
it work. With real source, this could have been translated into
free help.
A product with promise is a boat anchor. Anyone with sense has given
up on it.
The price and raw hardware specs were quite good. The software and
support were useless. I'm embarassed that they used my software.
================
The OpenWRT project is great. It appears to have too much work for
too little manpower. So join in!
--
The Toronto Linux Users Group. Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists
More information about the Legacy
mailing list