[OpenWrt-Devel] Introducing the LEDE project

[Daniel Dickinson]

Sorry if this hits the list twice, I sent from the wrong email and got
bounced.

On 16-05-04 06:05 PM, Bob Call wrote:
> On Wed, 2016-05-04 at 21:19 +0100, tapper wrote:
>> On 04/05/2016 21:01, mbm wrote:
>>>
>>
[snip]
> 
> I'm kind of split on this issue because I run a faltering OpenWRT fork,
> feel that OpenWRT has grown beyond the scope of an "embedded" OS/distro

"embedded" is kind of an amorphous word.  Smartphone companies call
their devices embedded devices, and others would argue that unless are
as small as core Aduino you're not truly embedded (i.e. that if you can
run Linux you're already big and bloated).

In truth there is wide range of things than can be considered embedded,
and no single project can meet the needs/wants of all such devices.

I'm not sure what LEDE intends to make it's primary uses cases, and what
other uses cases will kept into consideration, but I find it unlikely to
be concerned with <4 MB flash and <16 MB RAM as that is too far from
where it is today (and for the types of uses cases I suspect it will aim
for).

It would be good to get clarity on what uses cases will be in play for
LEDE though.

> and the goals of LEDE don't fix the faults with OpenWRT or its
> community.
> 
> There are many issues that need to be addressed and maybe the only way
> that some could start the conversation was to start a fork. Sometimes,
> great things come out of forks and eventually make it back upstream
> (some cerowrt work comes to mind). Regardless of the reason for the
> fork, the community at large can start a conversation and work to
> resolve everyone's issues in a sane way.

I started a fork myself, although I've never announced it, as it's
nowhere near what I would consider a useful split, and likely will
remain primarily a staging tree for a larger project or projects (e.g.
openwrt and/or lede).

Of course I do have some use cases in mind that could make it useful for
purposes other than the sort of thing LEDE is interested in (e.g. VM/LXC
guest and hosts with minimal footprints), and which if I do get to
working on will probably result in enough divergence to make it require
a separate project.

> 
> I started my fork due to philosophical and technical reasons with the
> intent of address concerns that many in the free software community
> have, but during my work on this fork, I've found that:
> 
> * Many components in OpenWRT are becoming too bloated and make it
> difficult to use some lower-end routers/targets without neutering
> commonly wanted functionality. While this is not fully OpenWRT's fault,
> it would be good to start making an effort to get some upstream
> projects to work on reducing their footprint.

Unfortunately you're talking about not just lower-end but old, no longer
produced routers, at least in North America (where I live), and I've
noticed even from linux kernel and distros that old hardware dies from
lack of support even though it takes longer than for proprietary
solutions.  It's the old story of whether it's worth the time and effort
to squeeze the last bit of life out of the old jalopy or get a full
efficient compact.

The problem with technology is that keeping old hardware around is not
necessarily reasonable even for things like environmental reasons
because the technology is consistently improving in the areas we care
about over time; the ability of individuals to innovate and enhance
technology for the future is one of the reasons open source freedom matters.

Regards,

Daniel