Good qtns @Wash -shows you have read my post :-) Point1: Can we delink RTS from EVID?Yes we can. From the perspective of functionality, EVID does NOT need RTS to work and vice versa. However, from an audit perspective (e.g after potential rigging), the data in EVID can be compared to the data transmitted by RTS to identify ghost voters. Hence the symbiotic relationship - and the need for ensuring that at least one of these two systems MUST work. In cases where BOTH EVID and RTS fail, then voting exercise must be repeated since the exposure to rigging is far too high to the point where reliability of the results cannot be guaranteed as per the constitution (ref: secure, verifiable, transparent, accurate) In other words, we must accept that failure may occur, but we must also reject that failure will occur in totality. Where failure occurs in totality, we must be willing to repeat the process in the specific constituency/county/country. That is the cost of democracy. Point 2: Can EVID fail - Maliciously or otherwise?True you dont need 3G/Satellite systems for the EVID component since it is simply comparing your fingerprint at the point of voting, to the fingerprint you digitally supplied at the point of biometric voter registration (BVR). However, failure can occur in what we call false positives or false negatives arising from environmental conditions rather than technical (battery failure etc). Humidity, dust, temperature variations can make the fingerprint reader fail to recognise you (false negative). More advanced EVID can then identify you by your eyes (Iris), voice or other biometrics but I doubt the IEBC system will be this advanced. (btw: This is what CS Mucheru should have been talking about, not Alshabaabs ;-) Solution: allow for manual identification - ON CONDITION that the RTS will work within reasonable time . We must set time frames within which RTS should work, it cannot be acceptable that it works, 1, 2 or 3days later since this defeats the whole essence of RTS, which is to provide randomness in the elections results; making it difficult for parties to do the mathematics needed to pass the magic 50%+1. This mathematics only happens when their is a pause or stability in the results being released. Point3: CORD insisting on pure electronic system vs Jubilee insisting on manual system.Whereas electronic systems reduces chances of manipulation, CORD stubbornly refuses to appreciate Point2 above, which is that e-systems can fail. Jubilee on the other hand stubbornly refuses to accept that manual systems introduced are open to manipulation. There must be a middle ground that allows for failure, without necessarily allowing for rigging. Both sides do not want to spend time on that agenda. How to plan for failure, while reducing chances of rigging. Point4: VSAT technology?Ofcourse VSAT can work. Even if Alshabaab blew up the GSM masts, we can still transmit results via VSAT/satellite phones. Last time I checked, terrorists do not yet have technology or grenades that can blow out and bring down satellites from the skies. walu. From: Odhiambo Washington via kictanet <kictanet@lists.kictanet.or.ke> To: jwalu@yahoo.com Cc: Odhiambo Washington <odhiambo@gmail.com> Sent: Friday, December 30, 2016 9:14 PM Subject: Re: [kictanet] Manual Backup and Elections 2017: Is the CS ICT Honest Walu, I read your views, but I have one or two observations:You say "Conversely, if the RTS fails but the BVR and EVID work perfectly, there should be less cause for alarm. Essentially, the three subsystems have a symbiotic relationship that can be used to validate or cross-check each other." => I do not clearly get the symbiotic relationship between the three systems as far as the main issues of contention (EVID) are concerned.If the systems are interconnected, then, looking at it from an SQL perspective, the RTS system borrows only one column from the BVR tables - Total Registered Voters in a particular Constituency, which I believe is just a factor for cross-checking the results (reminds me of Tiaty saga).However, this isn't necessarily part of the critical system that is supposed to stop dead voters from resurrecting and voting! I therefore think that we can mentally (or even practically de-link the RTS from the EVID to stop this insistence on connectivity, which gives birth to the "manual backup", no? You say "Sometime the failure is maliciously engineered, while other times it is a reflection of the genuine weakness inherent within man-made systems." ==> With specific reference to EVID, I am of the opinion that it is pretty easy to mitigate failure of the system by having 3 sets of the system, which is affordable. That would address the "technical failure", but not a human/maliciously engineered failure, because the humans can kill the three or even all of the equipment. If they do this (corrupt the static DB - as that is the only show-stopper), then really, there should be no voting. I hope that doesn't happen. We still don't need 3G/4G/VSAT for this. You say "So Cord, just as prescribed for Jubilee, should be discussing what level of electronic failure is acceptable, beyond which the results can no longer be acceptable given the potential exposure to manipulation that would arise from the manual alternatives." ==> Jubilee are advancing/contemplating the imaginary failure of connectivity occasioned either by absence/failure of fast network (3G/VSAT) or Al Shaabab knocking off what is there. This is more like making a nightmare a reality instead of the dream that it is. CORD is insisting that EVID should be used without the option of the "manual backup" and we all know that EVID doesn't require this connectivity, which supports the CORD argument. You say "On the other hand, electronic voter identification (EVID) and the results transmission system (RTS) are quite time-sensitive. If they failed, manual intervention may be the only option available" ==> I still insist that EVID has very little to do with RTS. EVID is being used statically. The equipment, at most, has the constituency register, not the whole national register, and at the least, has just the registration/polling centre register. RTS is a system that kicks in later, once EVID has completed its role. RTS waits for data from humans - clerks/agents/presiding/returning officers. They can all congregate at the County HQ and send this data. Most County HQ have 3G. If they don't, VSAT is something that can be set up in less than 1 hour! Your other view go well with issue about addressing possible failures, but in no way support the "Manual Backup" for EVID. This manual backup thing is a red herring, visible immediately you de-link connectivity from the debate. Which brings me to your conclusion: "So who is right and who is wrong? Unfortunately, both sides are right and at the same time, very wrong." ==> That's because we are not there to talk to them and enlighten them. Hey, doesn't Jubilee/CORD have ICT experts in their stables/secretariats though?? :-) On 30 December 2016 at 19:55, Walubengo J via kictanet <kictanet@lists.kictanet.or.ke> wrote: @Barrack, My views and solutions were shared ealier. But just incase listers did not read or have questions, you can find them here. WALUBENGO: On electronic polling, both Cord and Jubilee are | | | | | | | | | | | WALUBENGO: On electronic polling, both Cord and Jubilee are Nothing stops the opposition from abusing a failed system, but the incumbent always has an upper hand | | | | walu. From: Barrack Otieno via kictanet <kictanet@lists.kictanet.or.ke > To: jwalu@yahoo.com Cc: Barrack Otieno <otieno.barrack@gmail.com>; JImmy Gitonga <jimmygitts@gmail.com> Sent: Friday, December 30, 2016 6:26 PM Subject: Re: [kictanet] Manual Backup and Elections 2017: Is the CS ICT Honest Dear all, In the meantime, we encourage all of you to raise the substantive issues you have the way Washington and other colleagues have done. We have started collating the views for submission to the Senate next week. Best Regards and wishes for the new year On 12/30/16, Collins Areba via kictanet <kictanet@lists.kictanet.or.ke > wrote:
Let's dissect the problem into pieces. 1: voter registration: collecting details, photos and fingerprints. (Multiple data types)
2: verification: ascertaining that registered persons are in the system, and dead / expired ones are removed from the system. (Boolean: yes / No)
3: voting: choosing from one of several options.
4: tallying : counting the choices at the polling stations and recording the results on paper or device.
5: transmission: sending this information to regional and national tallying centers.
Maybe the good CS can explain how al shabbat can disable IT solutions so much that manual "backups" would suffice.
On 30 Dec 2016 17:58, "Grace Mutung'u via kictanet" < kictanet@lists.kictanet.or.ke> wrote:
Thank you Wash for initiating the discussion. I also wondered whether a complimentary system was used in voter registration and where this system resides.
I remember a quote by the IEBC CEO during the Kenya IGF where he stated that being a Republic based on democracy, we have made elections the only means to access power. https://livestream.com/ internetsociety2/kigf He recalled the use of tech in the 2010 Referendum, 2013 elections and the various by-elections that have taken place. In the Referendum and most by-elections, there was not much contest about use of technology while for 2013 some issues were raised- these included multiple registers, voter impersonation and transparency. The tech community has an important role to play in demystifying some of these concepts. a) The wording of the amendment read "complimentary mechanism for identification of voters". It has now been expanded to include transmission of election results "where technology deployed initially fails". What would this mean, in the case of identification of voters and in the case of transmission of results? What complimentary systems were envisaged here? "Manual backup?" The ambiguity in the wording is a challenge as it leaves too room for interpretation in a system of high contests.
b) The mischief that technology was meant to cure in elections management was among others allegations of voter impersonation and transparency in management of elections. Tech is therefore supposed to achieve simplicity, accuracy, verifiablilty, security, accountability and transparency. Is the conversation about a " complimentary" system a necessary one at this stage?
Outside of the amendment, has anyone come across the data that CA presented on network coverage in the counties? A visualisation of that data besides the polling stations would be useful in helping us identify the specific polling stations/tallying centres that are not covered. I am asking this because the presenters spoke of areas not covered by network as opposed to polling stations/tallying centres not covered.
Raha tupate na ustawi
2016-12-30 13:54 GMT+03:00 Victor Kapiyo via kictanet < kictanet@lists.kictanet.or.ke> :
I agree. We should put together our submissions as the ICT community and present the same to bunge.
Victor
On 30 Dec 2016 13:50, "Dorcas Muthoni via kictanet" < kictanet@lists.kictanet.or.ke> wrote:
Thanks Walu, it's time for us to stand up. Let's demystify this tech.
On Dec 30, 2016 1:43 PM, "Walubengo J via kictanet" < kictanet@lists.kictanet.or.ke> wrote:
I think this is an opportunity for the ICT fraternity to take up the challenge and demystify electronic systems in elections. I believe this forum has the most neutral platform where the media, academia, operators, regulators, government, legal and other interested parties can brainstorm on this.
I propose that early in the year, a face-2-face roundtable TV /Radio broadcast (NTV, Citizen, KTN?) deliberation to break this down -perhaps at Strath University (CPIT are you there?).
A lot has been written on the issue of electronic systems in elections but seems nobody READS, least of all politicians from both sides of the divide. I can imagine a cast of the following:
IEBC: CEO or Rep? Regulator: CEO or Rep? Operator: Safcom/Airtel/Telkom? ICT Min: Minister or rep? Academia: MMU/Strath/UoN/? Political Party: Jubilee+CORD Rep? Moderator &Broadcaster: Media (Citizen, NTV,KTN) Convenor: KICTAnet -GG are you back from holiday? Sponsors: Anyone?
If we do not hijack this ICT conversation, the politicians will run with it in the wrong direction and we might find ourselves exactly where we were in 2007.
walu.
------------------------------ *From:* JImmy Gitonga via kictanet <kictanet@lists.kictanet.or.ke > *To:* jwalu@yahoo.com *Cc:* JImmy Gitonga <jimmygitts@gmail.com> *Sent:* Friday, December 30, 2016 1:14 PM *Subject:* Re: [kictanet] Manual Backup and Elections 2017: Is the CS ICT Honest
Thank you Odhiambo Washington,
I have the same concerns myself. I reached the conclusion that it would be nice if "ICT Experts" could lay their hands on a BVI machine as well as a and show the rest of us what the problem really is. The ERT issue is a red herring. It has worked flawlessly in the bi-elections that have happened ever since. With PKI and 2 factor authentication, this can be solved for election day.
I am sure Victor Kyalo and Joseph Mucheru could make this possible. Call it a "Kenyans as ICT stakeholders" meeting. All Listers with time will begin to be asked by their family or neighbours, what the issue really is. I, for one, do not want to echo the CS's words.
I think the CS and the PS should help us help them. Otherwise they will be on their own when the political vultures come calling.
Regards, Jimmy Gitonga
Message: 4 Date: Fri, 30 Dec 2016 11:30:08 +0300 From: Odhiambo Washington <odhiambo@gmail.com> To: Kictanet <kictanet@lists.kictanet.or.ke > Subject: [kictanet] Manual Backup and Elections 2017: Is the CS ICT Honest Message-ID: <CAAdA2WPFoRvjF5Bodk+sHb-P4_ rUp2AXpA9Q3zAk-UR57cqwGw@mail. gmail.com <CAAdA2WPFoRvjF5Bodk%2BsHb-P4_ rUp2AXpA9Q3zAk-UR57cqwGw@mail. gmail.com>> Content-Type: text/plain; charset="utf-8"
Dear Listers,
I am at that position where I feel very lost. In fact, I feel like I am quite detached from the reality.
All along, I have keenly considered this matter that seems to have divided the country down the middle: *Manual Backup* during the 2017 voting process. From the Jubilee/govt side this is a do or die and so it must be there. This govt side seems hell-bent on confusing the masses, as well as the experts like the ICT Community. From the Opposition side, the agenda seems to be very clear - that of ensuring integrity of the Voters Register and stopping 'ghost voters' from ever casting their votes.
That brings us to the famous acronyms - BVI (Biometric Voter Register) / BVI (Biometric Voter Identification). Having been to a Voter Registration Centre (later to become a Polling Station) to register as a voter, I did look at the equipment in use for the registration process. I saw the laptop which was fitted with a camera and fingerprints scanner. All these require power to run. I did not dwell on how they were powered, but probably there was a battery backup somewhere (besides the electricity) given that they needed to run for a whole day for several days during the voter registration process. When it comes to the Elections, they only need to run for about 11 hours. My point here is that of *Backup Power* should it be that there's electricity blackout and the built-in batteries can't last the whole day. That backup is very important.
However, I did not see any piece of equipment which could suggest that the equipment in use required any form of connectivity back to some central server in order to function! Which now brings me to the currently national debate - Manual Backup during the Poll Day. What is it? Was the CS honest with his presentation before the Senate/Amos Wako committee yesterday? Does the CS himself really believe in the content of his presentation? I am asking that because I watched him and I don't believe him. I actually think he mislead the committee, and hence the nation at large.
Someone please prove me wrong. I am at that point where I believe that the BVR/BVI does NOT require any form of connectivity and so this Manual Backup being touted by the ruling coalition side, strongly supported by the ICT CS is a big lie. WHY?
My very first answer: Simply put, *when there was no requirement for a manual backup during voter registration, it goes without saying that there is NO NEED for on the polling day.*
1. For the issue that is in contention - BVR (used for BVI during polling) - this is a database that can be (and should be) statically stored on the equipment for each polling station. We are not supposed to rely on the Mobile Network to access this voters database. And every polling station can have two/three laptops/Biometrics scanner/Backup batteries to ensure that the voter identification doesn't fail. Some excuse has been fronted about some voters being mechanics, such that their fingerprints wouldn't be recognized by the BVI systems hence need for manual identification. My take on that is that every voter must carry their voter's card on that day. The clerks can check that card number against the electronic system - enter it, and it will bring the person's picture, ID number, etc and let him cast his ballot.
2. For electronics results transmission (ERT), this is not even necessary in the first place. We can have the results collated/announced at the tallying centres after being certified - forms 36A, and such. However, if the ERT must be done, the data footprint is so tiny that a 2G network can be used. Besides, it can be an SMS based system, which doesn't require 3G or VSAT. The results data isn't that large - it can't be in Megabytes to be sincere. Well, VSAT can be used if they MUST, but this is after the voting process itself is complete, has nothing to do with BVI.
The ERT and the BVR/BVI are two distinct systems. That is what I want to believe. The ERT gets feedback from a manual process - of voters casting their vote, clerks/agents counting, verifying, and certifying, filling requisite forms then communicating the same via some customized phones which are programmed to communicate to a backend system. Am I right on that??
Now the big question here is, where do we need this much touted manual backup where network connectivity is being used as the major reason???
(a) Citing terrorism and the possibility of Al Shabaab knocking off base stations seems like well thought out lie meant to cover our eyes! If they attacked an area, I doubt there will be voting in the 1st place. And even so, the network connectivity is not required for BVI so there is no disenfranchising anyone if there is no manual backup (whatever that is).
(b) Citing hacking is neither here nor there for a BVR/BVI system because it's not being accessed live during the voting. It's a static database, unique to the polling station, resident on the laptop used by the clerks. The only hacking that can be done then can only be by an "insider". Quoting Victor Kapiyo from Social Media, "*I guess it's a question of trust. Trust in systems and in trustworthy people to do the right thing. For M-Pesa, or KCSE results, we trust both. For IEBC, I guess the jury is still out*."
The main issue is not allowing the dead voters to rise again to vote in the presidential vote, then disappear. So the important component here is the BVR/BVI, the credibility of the register and hence the vote.
At what point does the BVI system require this connectivity they are talking about, which then necessitates the so called "manual backup"?
Did the CS ICT lie to the Senate?? Did the CAK lie to the Senate in supporting the lie from the CS??
There is insincerity in this whole debate about 'manual backup' and the ICT community seems to either support it or is simply lost in the pool of confusion being peddled by politicians.
-- Best regards, Odhiambo WASHINGTON, Nairobi,KE +254 7 3200 0004/+254 7 2274 3223 "Oh, the cruft."