Re: [Legba2013] PapaLegba 2013 review.

[Peter S. <pe...@st...>]

tim panton wrote:
> What didn't work?
> 1) We set up late.

If Papa Legba is to run a useful cell network without early arrival
then I think it is an absolute requirement to arrive better prepared.
No building on playa. Build in advance and just deploy on playa.

That means e.g. mounting antennae to masts before loading masts
into truck and loading equipment into truck in a manner optimized
for deployment rather than for space or time, which surely requires
both more space and more time.

Better preparation also means a clear distinction between pre-event
participation and during-event participation. Many things are only
relevant pre-event and if all preparations are to take place in a
single location then anyone who is not at that location will be
unable to participate in any and all pre-event matters.


> 3) We lost power _every_ night.

This had very significant impact on the radio network. I had hoped
that my request for running all infrastructure off 12V DC would have
meant that we ran more equipment directly off batteries this year,
but we weren't. I did however also not look into the power situation
too closely, neither in advance nor on location.


> It is frankly amazing that anything worked at all.

Many good people put in much hard work and together I think we made
the best we could out of the circumstances.

What amazes me is that performance was *so* much worse than last year.

I'm not sure that more hands on playa would have made a big difference.


> 6) We broke our promise to Ice to deliver a gate SMS application
>    through to monday evening.

I hope that Papa Legba will continue to communicate with Ice.


> 7) we didn't communicate our purpose very well to other burners.

Personally I think I successfully communicated our purpose to those I
spoke with - I got a lot of good response and none bad, and I also
really liked our banners on the front porch tent and on the truck!

But as a theme camp I think Papa Legba could do much better. However -
given that we didn't offer very good user experience in the cell
network it is just as good that we weren't heard so broadly.

Most other burners who came to see us were simply interested
in briefly using a phone, any phone, or an internet connection.

We would diligently explain to them why we were really there,
to build a cell phone network that didn't work for Verizon or AT&T
customers, while they waited for their turn to use the hardphone
and listened to us politely, but with varying degrees of genuine
interest.

They just needed to communicate right then and there. I'm glad and
proud that we made that possible for so many who came to see us.


> What worked?

I think the AT&T opt-in ritual worked late in the week. Nice!


> 1) It looks like the new L3 held up ok, although I have no stats on that.
> 5) The inter-base station 3.5GHz network

Are these two one and the same? The 3.65 GHz connectivity was good,
excellent link quality in undisturbed spectrum. The radios had the
factory default insecure configuration, which could have interfered
with the NOC provisioning system (yes, even though it's a different
band - no, not RF wise but end-user wise) but that was quick and easy
to address.

NOC told me that they will likely use 3.65 GHz for some links next
year, so it will be important to coordinate with them (probably at
the WiFi summit) to avoid RF interference.


> 6) PlayaNet as backhaul worked ok for us this year - perhaps
>    becuase we didn't require anything special of it.

I'd prefer to call the NOC network uplink. As uplink it almost always
worked well, once it had been configured, precisely because we didn't
use it as backhaul, ie. we didn't require anything special - no manual
configuration by NOC and no special quality of service for core
signalling.


> 7) The provision of wifi and a hardphone was popular again.

I would even go so far as to say that that single hardphone was our
primary contribution this year - but that was plenty for those who used it!

Hermione made an interview with one guy, I hope it is published soon.

It also made me very happy that people came to us because they had
learned from others that we might help them communicate.


> 8) group sms

I'm sure that the group SMS functionality was solid in the core, but
the fact that my phone was never registered when I looked at it (it
was powered on for the duration of the entire event), that I couldn't
register manually more than a few times in four days, and that I at
one point had received the same SMS 8 times all in all meant that I
perceived group SMS as completely unusable.


> Overall, I'm doubtful I'll do this again, it will be difficult to
> justify the time spent for the results unless we make significant
> changes in our approach.

I feel similarly. For me it's critical that each iteration brings
significant progress along some axis - doing better, smarter, more
robust, simpler, easier, more efficient, faster or in some other
way constantly improving. Anything less is just a waste of life.

Finally, communication is the foundation of all team work, which
is what I believe accelerates progress the most.

That's also why our gift of communication is so valuable.


//Peter