On Thursday 19 July 2007 18:11:55 Harry Lachanas wrote:
> Imagine this scenario,
> you have LEAF boxes spread all over your country,
> would you trust floppy disks on your installations even with backups
> around?
>
> Floppy disk devices have movable parts, CFs don't, usb-sticks don't,
> I personally haven't used any floppies for 4 years now, period.

Harry;

I agree  - the less movable parts a router has, the more reliable it will be 
over the years.

Therefor LEAF provides ISO images, USB images and the ability to boot from 
harddisk, CF and whatever.

One good reason to provide a floppy-based version is not just about to reuse 
old hardware, which btw is always a good starting point for new users to get 
in touch with a more secure and better adaptable software than delivered with 
the usual SoHo routers. For developers the "floppy orientation" is more or 
less a synonym for discpline and concentration on the goals. 
In contrast to the early days the costs for storage space is neglectable 
today. And yes, USB sticks are cheaper and easier to buy than floppy disks. 
It's easy to build a router based on Debian or any other Linux distro. But 
it's questionable if users have that easy that much control over the software 
installed, and therefor the potential security issues.
In the long run the "embedded"  aspect in LEAF may get more attention than 
today, so space will become more important again.

The whole issue started with the question "can we go with a 2.6 kernel and 
what about the floppies then?".

IMHO moving forward to a 2.6 kernel is a good idea. 
For the floppy versions the current branch can be maintained for a long 
time  - as long as users understand that only harwdare supported by the 2.4 
kernels can be used and that not every new package (version) can be added, 
because it can't be backported (easily).

A floppy version based on 2.6 kernel may be doable as well, but for less usage 
scenarios than today (dhcpd and pppoe are both on the current images), maybe 
without a ntp daemon and other software. So either as proof-of-concept or by 
building several floppy images instead of one as today.

So it's mainly not a question of preserving floppies as one target or forget 
about it; it's mainly a question of manpower to start with new versions, 
probably ports to other architectures beyond the x86 as of today and to 
improve the tools we developed over the last years to build LEAF software.

kp  

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
------------------------------------------------------------------------
leaf-user mailing list: leaf-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/leaf-user
Support Request -- http://leaf-project.org/

Reply via email to