Eircode - its implemetation (merged)

plodder

clewbays wrote: »

Are you saying 80% to 90% of cases reported included an Eircode or do you mean things like a street name, or a park name etc. because it is very hard to understand how even 80% of incidents would be directly linked to a postal address?

Looking at the slide, I'd estimate that about a third of the categories could be assumed to always link to an Eircode. For the others, some of the time they might, but in general they wouldn't.

PDVerse

plodder wrote: »

You are exaggerating the benefit of this "feature".

a) Would it really be such a problem if the emergency services arrive at a location 100 metres away?

Are you seriously asking this question? Have we reached the "jump the shark" moment of anti-Eircode bias where an argument can be made that ambulances turning up at the wrong location isn't a big deal?

plodder wrote: »

b) The benefit only applies to non unique addresses and the generally accepted figure for these is 35%. It does not apply to urban areas where addresses are unique and eircodes can be used to validate addresses without error. Why impose the limited advantage of this feature on the 65% majority? Why not have them random in rural areas and sequential in urban areas?

Yes, it is more beneficial to all of rural Ireland. Non unique addresses were a priority of the design.

I always recognise the "if you make a small mistake with a hierarchical code you will be close to your intended destination" argument as an indicator of bias. You have to be completely biased to assume that the error in the Eircode will be in the last or second last character rather than the fourth or fifth character. How far away will you be then?

I do not think a design feature that assists the National Ambulance Service to respond to an emergency faster is a "limited advantage".

plodder

PDVerse wrote: »

Are you seriously asking this question? Have we reached the "jump the shark" moment of anti-Eircode bias where an argument can be made that ambulances turning up at the wrong location isn't a big deal?

I related an example a while ago of where I found an ambulance about a mile away from its destination. Had they been within 100metres there wouldn't be a problem as they'd have seen someone frantically waving from the front door.

You have to weigh up the supposed advantages of this system with the difficulty in remembering essentially random codes. And that's before we get into the whole proprietary database vs geocode debate.

Yes, it is more beneficial to all of rural Ireland. Non unique addresses were a priority of the design.

Again. Why implement it across the spectrum and not just in rural areas where the benefit was to be had?

I always recognise the "if you make a small mistake with a hierarchical code you will be close to your intended destination" argument as an indicator of bias. You have to be completely biased to assume that the error in the Eircode will be in the last or second last character rather than the fourth or fifth character. How far away will you be then?

We're talking about Eircode here, not some hypothetical hierarchical code and the random part of Eircode, is where these errors could occur

I do not think a design feature that assistes the National Ambulance Service to respond to an emergency faster is a "limited advantage".

It would be interesting to find out what percentage of calls NAS gets with Eircodes. If they were easier to remember then I suspect the number would be higher.

PDVerse

plodder wrote: »

I related an example a while ago of where I found an ambulance about a mile away from its destination. Had they been within 100metres there wouldn't be a problem as they'd have seen someone frantically waving from the front door.

Bias Alert: Because obviously every location within 100m of a house is in direct line of sight.

plodder wrote: »

You have to weigh up the supposed advantages of this system with the difficulty in remembering essentially random codes. And that's before we get into the whole proprietary database vs geocode debate.

Again. Why implement it across the spectrum and not just in rural areas where the benefit was to be had?

You have to remember one code, no more. The first three characters are the same as your neighbours. That leaves 4 characters. Remembering 4 unique characters is not a difficult challenge, so yes the advantages far outweigh the "alleged" disadvantages.

plodder wrote: »

We're talking about Eircode here, not some hypothetical hierarchical code and the random part of Eircode, is where these errors could occur

You don't appear to understand the logical conclusion of your own argument.

plodder wrote: »

It would be interesting to find out what percentage of calls NAS gets with Eircodes. If they were easier to remember then I suspect the number would be higher.

This isn't 1972. Eircodes are entered on websites, provided to people who put them into CRM systems etc. The IT systems translate the data into usable, actionable data suitable for each use. For most private citizens that's simply entering into Google Maps and getting directions. For businesses it can be assigning to the correct Engineering area or Sales Territory, or Delivery Depot. None of that is done by staring at the code.

We're in the implementation thread, and I can state from my experience that none of the implementations I've been involved with have raised this as an issue. SUSI had a 92% Eircode rate on their forms (yes they were validated, no, people didn't just enter anything) so in the real world the last four characters of the Eircode do not present an issue.

When evidence challenges your assumptions you are meant to change your mind.

oscarBravo

plodder wrote: »

...the difficulty in remembering essentially random codes....

...is a fiction. Mobile phone numbers are essentially random codes. People remember their phone numbers.

GJG

plodder wrote: »

You are exaggerating the benefit of this "feature".

a) Would it really be such a problem if the emergency services arrive at a location 100 metres away?

Yes, a miss is as good as a mile if you are dying of a heart attack at a house 100m away from the ambulance who have no way to locate you. But look at the addresses, like SHANKILL, CAVAN. That is more than 100m across. I am personally familiar with warrens of rural roads, several kms in diameter, all with identical addresses.

Are you seriously suggesting that, in addition, giving them almost-identical Eircodes would have been a good idea?

plodder wrote: »

b) The benefit only applies to non unique addresses and the generally accepted figure for these is 35%. It does not apply to urban areas where addresses are unique and eircodes can be used to validate addresses without error.

That's flat wrong. It depends on what you call non-unique, the 35 per cent refers to absolutely no difference between the addresses; but there is still plenty of scope for confusion between 110 High Street and 101 High Street.

In many cities there are streets officially called Upper / Lower / East / West / North / South / Great / Little (say) George's Street, all of which are referred to popularly just as George's Street. Eircode is set up to make sure that none of those have easily-confused codes. If you genuinely can't see the point in that, there isn't much more to say.

plodder wrote: »

Why impose the limited advantage of this feature on the 65% majority? Why not have them random in rural areas and sequential in urban areas?

You are imagining that it is even possible to keep codes sequential. What happens when a building is demolished and an apartment block is built in its place? You would quickly end up with the most confusing option, a partially sequential code. Have you ever tried to find a house on a street where some houses don't yet use the 'new' numbers, introduced 50 years ago or more? And what do you do when a rural area is urbanised?

And if you could overcome all of these issues of such a dog's dinner, and cope with confusion of multiple overlapping systems, why would you bother? What possible benefit are you pursuing?

GJG

plodder wrote: »

You have to weigh up the supposed advantages of this system with the difficulty in remembering essentially random codes. And that's before we get into the whole proprietary database vs geocode debate.

What on Earth are you talking about?

My neighbour's phone number is totally different to mine. Are you seriously saying that some magical force would help me remember my phone number better if it were in sequence with hers?

plodder

PDVerse wrote: »

Bias Alert: Because obviously every location within 100m of a house is in direct line of sight.


You have to remember one code, no more. The first three characters are the same as your neighbours. That leaves 4 characters. Remembering 4 unique characters is not a difficult challenge, so yes the advantages far outweigh the "alleged" disadvantages.

Bias Alert.

It doesn't really further the discussion much by saying that, does it?

What if the whole code bore a direct relationship to your neighbour's? Would that not make it easier to remember again?

You don't appear to understand the logical conclusion of your own argument.

No idea what your point is.

This isn't 1972. Eircodes are entered on websites, provided to people who put them into CRM systems etc. The IT systems translate the data into usable, actionable data suitable for each use. For most private citizens that's simply entering into Google Maps and getting directions. For businesses it can be assigning to the correct Engineering area or Sales Territory, or Delivery Depot. None of that is done by staring at the code.

Sure, and for most of the people at your conference, I'm sure Eircode is a genuine success, but it still depends on widespread public acceptance.

We're in the implementation thread, and I can state from my experience that none of the implementations I've been involved with have raised this as an issue. SUSI had a 92% Eircode rate on their forms (yes they were validated, no, people didn't just enter anything) so in the real world the last four characters of the Eircode do not present an issue.

When evidence challenges your assumptions you are meant to change your mind.

There's a big difference between the examples (NAS vs SUSI). For SUSI, you have the citizen with plenty of time to make the application, eager to be as compliant as possible, vs a panic stricken situation trying to recall information. They could hardly be more different.

plodder

Don't have time to deal with all of this, but I just want to make this point.

GJG wrote: »

That's flat wrong. It depends on what you call non-unique, the 35 per cent refers to absolutely no difference between the addresses; but there is still plenty of scope for confusion between 110 High Street and 101 High Street.

You don't seem to understand. These are unique addresses. The corresponding eircodes are also unique. You have two unique pieces of information to cross reference. It is highly unlikely that you would make the corresponding mistake in both pieces of information. I've explained this to you before.

PDVerse

plodder wrote: »

Don't have time to deal with all of this, but I just want to make this point.

You don't seem to understand. These are unique addresses. The corresponding eircodes are also unique. You have two unique pieces of information to cross reference. It is highly unlikely that you would make the corresponding mistake in both pieces of information. I've explained this to you before.

If you have an Eircode and and Address that are similar, how can you tell which one is correct? You can't. With the Eircode design you can detect if the Address or the Eircode is incorrect.

There is a reason why Emergency services implement a best practice policy of asking for related information independently. In high stress situations the person reporting their Eircode may agree that each of the addresses below are correct for their Eircode.

8 Silver Birches, Dunboyne, Meath
8 Silver Birches Crescent, Dunboyne, Meath
8 Silver Birches, Navan, Meath

I'm not going to debate exactly how often it might occur, because that's a ridiculous argument.

Not everyone implements best practice. If I'm a delivery driver or providing a service and I enter the Eircode at the end of the address it is highly likely I won't recognise the error as I'll be concentrating on the first part of the address which looks correct.

If I'm in a customer care call centre and I'm trying to take address information quickly I may call out just "ok that's 8 Silver Birches correct" to the caller to verify the address.

If I'm entering my Eircode on my mobile and it autofills the address I may not notice the address is incorrect because I'm concentrating on the first part of the address.

During testing we allocated Eircodes randomly and then tested how many similar addresses were also assigned similar Eircodes. The result was 7%

GJG

plodder wrote: »

Don't have time to deal with all of this, but I just want to make this point.

You don't seem to understand. These are unique addresses. The corresponding eircodes are also unique. You have two unique pieces of information to cross reference. It is highly unlikely that you would make the corresponding mistake in both pieces of information. I've explained this to you before.

As PD Verse pointed out, 'unique' addresses aren't always that unique.

I think you are misunderstanding how a disaster occurs. It's not something that happens when something goes wrong. Things go wrong all the time, in nuclear plants, aircraft, ambulance services. Mostly they have no consequences, because of redundancy.

A disaster is what happens when many things go wrong in very specific ways. And you are absolutely right that the scenario outlined is highly unlikely. All disasters are highly unlikely, but they still happen sometimes.

You seem to be hung up on some very personal way you would use to remember sequential codes. I've never heard of anyone else primarily using their neighbour's phone number or code or the like to triangulate and thereby remember their own. I suppose it might not be just you, but I'm willing to bet that it is so rare that it doesn't remotely justify removing a whole layer of redundancy that is important for avoiding confusion.

antoinolachtnai

GJG wrote: »

It's not clear what your issue is here, but the examples you give certainly back up my point. Sam was making the case for sequential codes, and I pointed out that even if it was possible (it's not) it would lead to properties with near-identical, or actually identical addresses having near-identical codes, and that would give obvious scope for confusion.

I pointed out that Eircodes are not assigned randomly; care is taken to make sure that addresses with similar Eircodes have starkly different addresses, and vice versa. The examples you give bear this out.

I am glad you took the time to look up all those addresses.

It just seems that there was a very complicated endeavour to do this, but it is by no means guaranteed to work (which eircode admits) and it causes a lot of problems. (There is no detailed documentation of the method to do this either.)

I don't see the great advantage of this, and as far as I can tell no detailed research that was done on it.

I'd be a lot more confident in the scheme if it were just published. It is hard to see why it isn't just published now, seeing as it is being given away largely for free (well, 50c/1000 requests, which is almost free) on Google anyway.

It is great to say that this is a better system than a normal postcode where each area has a code and each individual house has an individual code, but this wasn't evaluated in any depth and it certainly wasn't consulted upon.

If someone has a subjective view that it's a good system, then that's fine. But that's all it is, a subjective view. There is no point calling people 'biased' because they don't agree with a subjective view.

Where are you getting the figure of 16000? Are you suggesting that almost one per cent of the letters were misdelivered?

Yes. As i understand it, that is a pretty typical mis-delivery rate or non-delivery rate for fairly straightforward uniquely addressed mail. For non-unique addresses I'd expect the figure to be higher. The burden is really on eircode to produce figures and justify how they ensured codes were distributed correctly, not on me. But of course there are no figures.

Don't you think that the people getting them would have been alerted by someone else's name on the envelope?

What proportion of envelopes had the incorrect name on the envelope? What database was this name-picking exercise based on?And what about the problems with similar personal names in rural areas? What verification was carried out on the accuracy of the database or on the accuracy of the deliveries?

As far as I know, no one knows the answer to any of these questions.

And in any case, it is trivially easy to verify it on the Eircode map.

It actually isn't. It is easy if a person has Internet access in their rural abode, knows how to work a computer well, and knows how to read a map. These are are not the sorts of people who typically have trouble directing an ambulance.

sondagefaux

clewbays wrote: »

sondagefaux wrote: »

And finally, how Waterford Council is implementing Eircodes:

... 8o-90% of cases reported to the council have a location component.

Are you saying 80% to 90% of cases reported included an Eircode or do you mean things like a street name, or a park name etc. because it is very hard to understand how even 80% of incidents would be directly linked to a postal address?

I'm not saying anything. I quoted a tweet concerning a presentation given by a representative of Waterford Council, who said that 80% - 90% of issues reported to them had a location component. Presumably they can ask for Eircodes when someone reports an issue at or near their address, or get it from a database when the person gives their address.

sondagefaux

a) Would it really be such a problem if the emergency services arrive at a location 100 metres away?


If my child is trapped under a tractor, I don't want the emergency services arriving at my neighbour's house, getting out of their vehicles, then back in again once they've realised they're at the wrong location, before getting to my house.

What happens if adjoining sequential Eircodes were separated by a physical barrier, like bollards or some other barrier dividing up a street?

The street I live on is partly pedestrianised and there are bollards at either end of the pedestrianised section.

If people try to get to my house simply using the UK postcode, they often either end up on the main road around the corner, or at the other end of the same street but blocked by the bollards on the pedestrianised section from driving up the street to my house. Getting around the bollards requires them to drive out of the other end of my street, turn right, then right again onto the main road, then up the main road for a couple of hundred metres, then right again into a side street that has a junction with the main road, and right again to get to the into the correct end of my street.

A friend who lives nearby has a similar problem. There are three house numbers, including his, that are out of sequence with the rest of the houses on the street. This means that his house, number 69, and two more houses, 70 and 71, are at the opposite end of the street from number 67 and 68, with the street divided by a small landscaped area with flower beds. If a delivery van goes to number 67, expecting to find 69 next door, they'll be at the wrong end of the street and unable to drive to his end without leaving the street, going around the corner and driving up the lane at the rear, then turning another corner to enter the street at his end.

b) The benefit only applies to non unique addresses and the generally accepted figure for these is 35%. It does not apply to urban areas where addresses are unique and eircodes can be used to validate addresses without error. Why impose the limited advantage of this feature on the 65% majority? Why not have them random in rural areas and sequential in urban areas?

There are two houses with the address: 2 Watergate Street, Bandon, Co. Cork. One example I have personal knowledge of which shows that not all urban addresses are unique. In any case, I can only imagine the howls of derision if it had been decided that Ireland should have two separate postcode systems.

TheChizler

antoinolachtnai wrote: »

It just seems that there was a very complicated endeavour to do this, but it is by no means guaranteed to work (which eircode admits) and it causes a lot of problems. (There is no detailed documentation of the method to do this either.)

A couple of weeks ago PDVerse posted a link to the document outlining in detail the how similar addresses were tested for. It's on the website.

antoinolachtnai

TheChizler wrote: »

A couple of weeks ago PDVerse posted a link to the document outlining in detail the how similar addresses were tested for. It's on the website.

You mean page 15 of https://www.eircode.ie/docs/default-source/Common/prepareyourbusinessforeircode-edition3published.pdf?sfvrsn=2 ? It is not detailed, and it is barely written in English.

I cannot see how the same thing would not have been achieved by simply having the last two letters of the code allocated non-deterministically.

The whole thing is a nice idea but there is no sign that it was really thought through.

(And at the same time, thousands of occupied homes and buildings don't have unique codes either, despite the value these unique codes would have for emergency services.)

plodder

GJG wrote: »

As PD Verse pointed out, 'unique' addresses aren't always that unique.

LOL. Come off it. You can't design a system based on nonsense like that. Take PDVerse's example

8 Silver Birches, Dunboyne, Meath
8 Silver Birches Crescent, Dunboyne, Meath
8 Silver Birches, Navan, Meath

These aren't even sequential addresses. Regardless of whether eircodes were done sequentially or not, these three addresses will have completely different, last four characters. The third one will have a different routing key. The chances of getting these mixed up is tiny. The addresses are unique. The codes are unique and they can be cross-referenced with a high level of accuracy. You can't escape that fact.

PDVerse

plodder wrote: »

LOL. Come off it. You can't design a system based on nonsense like that. Take PDVerse's example

These aren't even sequential addresses. Regardless of whether eircodes were done sequentially or not, these three addresses will have completely different, last four characters. The third one will have a different routing key. The chances of getting these mixed up is tiny. The addresses are unique. The codes are unique and they can be cross-referenced with a high level of accuracy. You can't escape that fact.

Plodder as you don't appear to understand the problem it isn't really surprising that you don't appreciate the solution.

PDVerse

antoinolachtnai wrote: »

You mean page 15 of https://www.eircode.ie/docs/default-source/Common/prepareyourbusinessforeircode-edition3published.pdf?sfvrsn=2 ? It is not detailed, and it is barely written in English.

I cannot see how the same thing would not have been achieved by simply having the last two letters of the code allocated non-deterministically.

The whole thing is a nice idea but there is no sign that it was really thought through.

(And at the same time, thousands of occupied homes and buildings don't have unique codes either, despite the value these unique codes would have for emergency services.)

Antoin, absence of evidence isn't evidence of absence. I wish I could purge my mind of the endless meetings that occurred during the two year competitive dialogue process. The Eircode database design is far, far more important than the Eircode design, but 80% of meeting time was devoted to the Eircode design. I do take offense when it is suggested that certain factors weren't considered or thought through. You may disagree with our design choices but please don't assert, without a shred of evidence, that we didn't go through a detailed and exhaustive design process.

antoinolachtnai

Then where has the documentation gone?

There was no sign of it when I FOI'd on November 11 2014 seeking

2. a copy of the high level design documentation including details of alternatives considered and rationales developed for the coding system ...

3. implementation plans for eircode

5. all records providing information about how the occupiers of addresses for which there is non-unique address might find out what their eircode is ...

- Capita is subject to a service contract and Capita records in relation to eircode are deemed to be DCENR records

A total of 16 documents is all that the search came up with. That is my evidence of absence. If you are telling me that there is other documentation which was not disclosed, I am very interested to hear.

I am sorry you take offence, but I am just going on the information I have been provided. Perhaps we have both been given wrong information somewhere?

I told you before about my experience of a member of the Postcode team engaging in deceitful behaviour. I took offence at that (but obviously I am not blaming you).

plodder

Ok, let's go through that example in more detail. Imagine the three eircodes are as shown. I don't know what the routing keys for Dunboyne, or Navan are, and they aren't easy to guess, but we'll just use the made up examples below:

8 Silver Birches, Dunboyne, Meath W50 XPA1
8 Silver Birches Crescent, Dunboyne, Meath W50 ZP12
8 Silver Birches, Navan, Meath W45 AB12

A 999 operator receives a call and asks for the address. The caller says
"8 Silver Birches Dunboyne". The operator types that in and W50 XPA1 comes up.

So, they ask what is your eircode? W50 ZP12 is the answer. But, that's not the right one.
The operator then says can you repeat your address please.

This time the caller says "8 Silver Birches Crescent, Dunboyne" Somehow the "crescent" got lost but now the address tallies with the eircode. We're good to go.....

The above works regardless of whether the eircodes are sequential or random, or carefully chosen by some obscure method. It makes no difference. The only problem would be if the following occurs:
- the caller calls in the wrong address
- the caller calls in the wrong eircode
- the wrong eircode happens to match the wrong address

That isn't very likely in my opinion. It's different obviously for non-unique addresses. In that case, it would be more likely that two sequential Eircodes could refer to the same address, and an error in either would not be detected.

PDVerse

antoinolachtnai wrote: »

Then where has the documentation gone?

There was no sign of it when I FOI'd on November 11 2014 seeking



A total of 16 documents is all that the search came up with. That is my evidence of absence. If you are telling me that there is other documentation which was not disclosed, I am very interested to hear.

I am sorry you take offence, but I am just going on the information I have been provided. Perhaps we have both been given wrong information somewhere?

I told you before about my experience of a member of the Postcode team engaging in deceitful behaviour. I took offence at that (but obviously I am not blaming you).

I don't have any issue with your concerns pre-launch that there was an adequate plan. Asking for documentation was healthy scepticism. There are lots of documentation not available to the department that wouldn't be available via FOI. It isn't because of an effort to hide anything, just the inner working documents of consortia, commercial agreements etc.

I appreciate your valid concern for dissemination errors/issues. Any estimates or assumptions made before the launch are now redundant. The dissemination has taken place. The majority have been checked on Finder (800k individual Eircodes were checked on day one of launch). They are in use by National Ambulance Service, delivery drivers and those providing services on a daily basis to a sufficient extent to expect that a significant error rate would have been reported by now.
There is an error rate, it isn't something that should be ignored, but there are insufficient reports of errors to warrant a widespread verification programme.

GJG

plodder wrote: »

That isn't very likely in my opinion. It's different obviously for non-unique addresses. In that case, it would be more likely that two sequential Eircodes could refer to the same address, and an error in either would not be detected.

I think we can agree, in that case, that the potential error is mostly concentrated in the 40 per cent of non-unique addresses. A matching double-error in a unique address, you are right, is very unlikely, although very unlikely things do happen sometimes and I wouldn't want to be the person explaining that to a grieving family.

The solution is therefore either Eircode as delivered, or your proposed dual system where unique and non-unique addresses have a different sort of postcode in order to accommodate your rather special way of memorising things. I'm happy for readers to decide for themselves which is most practical.

PDVerse wrote: »

I appreciate your valid concern for dissemination errors/issues. Any estimates or assumptions made before the launch are now redundant. The dissemination has taken place. The majority have been checked on Finder (800k individual Eircodes were checked on day one of launch). They are in use by National Ambulance Service, delivery drivers and those providing services on a daily basis to a sufficient extent to expect that a significant error rate would have been reported by now.

In addition, it's worth noting that the headcase on twitter has been loudly proclaiming that there are thousands of 'wrong' Eircodes, without even explaining what that might mean. I specifically asked for even one example and, despite several attempts, none were produced.

At some point the difference between the undetectable and non-existent becomes irrelevant.

NikoTopps

I'm sure it's been mentioned before but SIRO support eircode to check if your home is fibre enabled or not and it works very well!

sondagefaux

Eircode licensed by Dutch GSP and digital mapping company:
[font=Arial, sans-serif]dccae.gov.ie ‏[ltr]@Dept_CCAE[/ltr]  Dec 5[/font]
[font=Arial, sans-serif]Satnav provider @AND_Rotterdam has licensed #Eircode to accurately locate addresses used in their logistics & mobile mapping products[/font]

[font=Arial, sans-serif]https://www.and.com/digital-maps-2/[/font]

byrnefm

Didn't see this mentioned previously but Google Map searches for parts of an address now show the Eircode (in many cases) beneath each address, where a street has house numbers. Is this new?

byrnefm

I tried two Eircodes for apartments today on Google Maps (different parts of Dublin) .. and they now seem to be working! One was in D18 and the other in D04.

The bugs in relation to spaces still exist, in that I had to omit them or Google Maps returned a more 'local' Eircode routing key. Also, the apartment Eircode does not appear in the dropdown list of address 'selections' but otherwise, they both resolved to the correct location.

It's getting closer!

Any word on Garmin supporting them any time soon?

plodder

Just tried an apartment and it works. Also tried a business in a shopping centre and that worked too. Both in North County Dublin. The fact that business codes are working now is significant. There is a good reason now for retailers to use them.

GreenFolder2

It's all about getting it into the mapping systems. iOS native Apple Maps could be useful as would the various vehicle GPS platforms. It's an absolute pain in the rear entering Irish addresses on a typical car interface. Having Eircode rolled out would hopefully eventually trickle down onto new models and software updates.

oscarBravo

GreenFolder2 wrote: »

It's all about getting it into the mapping systems. iOS native Apple Maps could be useful as would the various vehicle GPS platforms. It's an absolute pain in the rear entering Irish addresses on a typical car interface. Having Eircode rolled out would hopefully eventually trickle down onto new models and software updates.

Android Auto - now available in standalone mode as well as working with compatible in-dash displays and aftermarket head units - aims to be mostly voice-controlled. I just tried asking it for directions to an eircode, and it worked.