Alternative design for Irish Postcode.

geographer52

MOD: See post 15

I have been watching this board for many years and have not felt able to comment, the reason for this was I was party to the bidding process, the group I was bidding with pulled out because we felt that the code structure mandated by DCENR was sub-optimal and hard to justify. In particular I felt that it was a missed opportunity to create the worlds best geo-coding system.

So how would I know? In the 1980's I ran the UK postcode system and I know what was done wrong in that design and implementation, since that time I have worked with standards bodies allover the world to help people to create and use addresses.

So what would a good Postcode System look like in Eire?

The first two Characters would match the car county code.

Then two numbers would match each significant Townland
(some Townlands are too small to be included and I know there would be rows and endless discussion, but the current code ignores Irish Geography and History so this would be better despite any inadequacy)

This would be followed by a second part, mixed alpha numeric that identified one side of a single road between junctions (Americans would call this a block face), a code would exist for every block face even if no properties existed on it.

A third part not general used would be a two character alpha numeric that identified a property on a block face. This would satisfy DNCER's requirement of identifying properties.

This code would be universal allowing a satnav to find every street segment in Eire, would allow geo-demographic analysis like the UK and Canadian codes. (As an aside the politicians who suggested that this kind of code generates a "postcode lottery" are confusing cause with effect, it allows government and citizen to see where there is existing inequality)

We tested this code against the whole road network in Eire and found that it was capable of recording all existing block faces and in the worst case allowed for 100% expansion , given that these would be new roads not properties it is difficult to envisage the circumstance in the next 50 years that would cause re-coding to be needed.

Finally it seems to me to be completely bizarre that this data and associated addresses are being sold at such a high price, the UK model. When the national benefit from free availability would be much greater, but then politicians the world over like to screw the last penny out of citizens.

clewbays

geographer52 wrote: »

I have been watching this board for many years and have not felt able to comment, the reason for this was I was party to the bidding process, the group I was bidding with pulled out because we felt that the code structure mandated by DCENR was sub-optimal and hard to justify. In particular I felt that it was a missed opportunity to create the worlds best geo-coding system.

Your preference as indicated is a hierarchical code - it is not making any use of GIS information and modern technology. Why didn't your group go with the LOC8 or TiCode approach? What would be World's best in your format?

The cost of ECAD is quite low for 1-2 users in an organisation buying from a value added reseller - a few hundred euro.

SPDUB

geographer52 wrote: »

The first two Characters would match the car county code.

What was your plan to deal with the Irish form of the county names since both Irish and English are allowed to be used on number plates .

Sam Russell

SPDUB wrote: »

What was your plan to deal with the Irish form of the county names since both Irish and English are allowed to be used on number plates .

Have you looked at number plates? The two letter (or one) are from the English language version of the name. The Irish language version of the name is written on the top of the number plate.

@geographer52 Is this a fixed length code?

I agree with your comments re charging for the use of the code. [Shhhh - don't tell Eircom or they will charge for the phone book! ]

oscarBravo

geographer52 wrote: »

I have been watching this board for many years and have not felt able to comment, the reason for this was I was party to the bidding process, the group I was bidding with pulled out because we felt that the code structure mandated by DCENR was sub-optimal and hard to justify. In particular I felt that it was a missed opportunity to create the worlds best geo-coding system.

So how would I know? In the 1980's I ran the UK postcode system and I know what was done wrong in that design and implementation, since that time I have worked with standards bodies allover the world to help people to create and use addresses.

So what would a good Postcode System look like in Eire?

The first two Characters would match the car county code.

Then two numbers would match each significant Townland
(some Townlands are too small to be included and I know there would be rows and endless discussion, but the current code ignores Irish Geography and History so this would be better despite any inadequacy)

This would be followed by a second part, mixed alpha numeric that identified one side of a single road between junctions (Americans would call this a block face), a code would exist for every block face even if no properties existed on it.

A third part not general used would be a two character alpha numeric that identified a property on a block face. This would satisfy DNCER's requirement of identifying properties.

This code would be universal allowing a satnav to find every street segment in Eire, would allow geo-demographic analysis like the UK and Canadian codes. (As an aside the politicians who suggested that this kind of code generates a "postcode lottery" are confusing cause with effect, it allows government and citizen to see where there is existing inequality)

We tested this code against the whole road network in Eire and found that it was capable of recording all existing block faces and in the worst case allowed for 100% expansion , given that these would be new roads not properties it is difficult to envisage the circumstance in the next 50 years that would cause re-coding to be needed.

Finally it seems to me to be completely bizarre that this data and associated addresses are being sold at such a high price, the UK model. When the national benefit from free availability would be much greater, but then politicians the world over like to screw the last penny out of citizens.

Sounds like an interesting design, with some caveats: first, not every county has a two-character car registration code. Second, does it preserve existing Dublin postcodes? Third, the house and the filling station visible here would have completely different postcodes, right from the very first character - not exactly helpful for organising deliveries. And finally, it's equally subject to two of the criticisms most often levelled at Eircode: it doesn't identify arbitrary (non-postal) locations, and it doesn't identify locations without a database lookup.

Sam Russell

oscarBravo wrote: »

Sounds like an interesting design, with some caveats: first, not every county has a two-character car registration code. Second, does it preserve existing Dublin postcodes? Third, the house and the filling station visible here would have completely different postcodes, right from the very first character - not exactly helpful for organising deliveries. And finally, it's equally subject to two of the criticisms most often levelled at Eircode: it doesn't identify arbitrary (non-postal) locations, and it doesn't identify locations without a database lookup.

I think the Dublin post codes could be accommodated by the two digits for the townland, which could incorporate Co. Dublin areas as well.

If it can identify every 'block face' or side of the road between junctions, then I think that covers every arbitrary place - not exactly but pretty close.

The database lookup would not be a problem if it was not proprietary and anyway does not go down to every property so it would be smaller.

The key question is whether it is fixed length or not.

oscarBravo

Sam Russell wrote: »

I think the Dublin post codes could be accommodated by the two digits for the townland, which could incorporate Co. Dublin areas as well.

Yes, they could. But were they?

If it can identify every 'block face' or side of the road between junctions, then I think that covers every arbitrary place - not exactly but pretty close.

Only if we're going to judge this code and Eircodes by completely different standards. What happened to the importance of having a postcode for every electricity pylon?

The database lookup would not be a problem if it was not proprietary and anyway does not go down to every property so it would be smaller.

If it doesn't include the exact location of every property, it's missing one of the single most useful aspects of Eircode.

What's a "not proprietary" database lookup?

Sam Russell

oscarBravo wrote: »

Yes, they could. But were they? Only if we're going to judge this code and Eircodes by completely different standards. What happened to the importance of having a postcode for every electricity pylon? If it doesn't include the exact location of every property, it's missing one of the single most useful aspects of Eircode.

What's a "not proprietary" database lookup?

This was a response to an outline of a possible design that was not pursued, so the details were not given.

The possible of identifying each and every letterbox was given, but no details of the actual design.

However, as has been pointed out, the design is now with Eircode (and all its failings) and there will be no other design for the foreseeable future.

oscarBravo

Sam Russell wrote: »

This was a response to an outline of a possible design that was not pursued, so the details were not given.

I know, that's why I was asking for them.

However, as has been pointed out, the design is now with Eircode (and all its failings) and there will be no other design for the foreseeable future.

That doesn't seem to have stopped anyone from criticising it... :pac:

Sam Russell

oscarBravo wrote: »

That doesn't seem to have stopped anyone from criticising it... :pac:

Well, criticism was due because it was an awful design. However it is now pointless as it will not improve with age but may gain acceptance (well it has to be accepted because it is the only postcode in town).

plodder

BowWow wrote: »

Don't understand why you are even discussing this? Postcodes are here, there wont be another system, the're being used, I'm using them every week in my business successfully, you cant reinvent the wheel......... move on...

We're stuck with Eircode clearly, but that doesn't mean it makes no sense to discuss its faults, or hypothetical alternatives.

oscarBravo

plodder wrote: »

We're stuck with Eircode clearly, but that doesn't mean it makes no sense to discuss its faults, or hypothetical alternatives.

Sure. It would be nice to discuss them honestly, though - and that has sometimes been missing from the debate.

OU812

geographer52 wrote: »

I have been watching this board for many years and have not felt able to comment, the reason for this was I was party to the bidding process, the group I was bidding with pulled out because we felt that the code structure mandated by DCENR was sub-optimal and hard to justify. In particular I felt that it was a missed opportunity to create the worlds best geo-coding system.

So how would I know? In the 1980's I ran the UK postcode system and I know what was done wrong in that design and implementation, since that time I have worked with standards bodies allover the world to help people to create and use addresses.

So what would a good Postcode System look like in Eire?

The first two Characters would match the car county code.

Then two numbers would match each significant Townland
(some Townlands are too small to be included and I know there would be rows and endless discussion, but the current code ignores Irish Geography and History so this would be better despite any inadequacy)

This would be followed by a second part, mixed alpha numeric that identified one side of a single road between junctions (Americans would call this a block face), a code would exist for every block face even if no properties existed on it.

A third part not general used would be a two character alpha numeric that identified a property on a block face. This would satisfy DNCER's requirement of identifying properties.

This code would be universal allowing a satnav to find every street segment in Eire, would allow geo-demographic analysis like the UK and Canadian codes. (As an aside the politicians who suggested that this kind of code generates a "postcode lottery" are confusing cause with effect, it allows government and citizen to see where there is existing inequality)

We tested this code against the whole road network in Eire and found that it was capable of recording all existing block faces and in the worst case allowed for 100% expansion , given that these would be new roads not properties it is difficult to envisage the circumstance in the next 50 years that would cause re-coding to be needed.

Finally it seems to me to be completely bizarre that this data and associated addresses are being sold at such a high price, the UK model. When the national benefit from free availability would be much greater, but then politicians the world over like to screw the last penny out of citizens.

Could the existing system be adapted to use this? The current random code would become hidden, the current routing key could be expanded & include other "non postal" areas & a first part could be added which would become the "Block Face"

marmurr1916

Sam Russell wrote: »

Well, criticism was due because it was an awful design. However it is now pointless as it will not improve with age but may gain acceptance (well it has to be accepted because it is the only postcode in town).

The only major criticism I've seen of it is that it's non-hierarchical, hardly relevant in an age of computer databases and mobile internet.

Apart from that, the fact is that it's a postcode, intended to identify postal addresses, not a geocode.

Sam Russell

marmurr1916 wrote: »

The only major criticism I've seen of it is that it's non-hierarchical, hardly relevant in an age of computer databases and mobile internet.

Apart from that, the fact is that it's a postcode, intended to identify postal addresses, not a geocode.

There are several criticisms.

1. It is based on routing codes that have no relevance on the ground. Some are non-contiguous, and many cross county boundaries.

2. The routing codes are over large in some areas and quite small in others.

3. The routing codes are based on a subset of An Post's postal towns. They have about 2,000 of them but have reduced them to about 100.

4. The retention of Dublin post codes was part of the spec but should have been incorporated differently as it has led to too large an area for each single routing code.

5. The random element is just daft.

6. It appears to be designed to maximise its use for monetizing the database.

Although the code is designed as a postcode, it could have been the basis for many other uses, such as navigation, statistical analysis, etc. However, we dumped the PayPars and eVoting, and we are in the process of joining the two Luas lines, so we will find a way of living with Eircode - but it could have been so much more. Remember, the Harcourt St line was closed by Todd Andrews for political/personal reasons, and has cost the state a huge amount to replace it 50 years later.

Eircode did not need to be based on An Post's method of handling post because they are fully computerised and fully automated, so any code would do.

oscarBravo

Sam Russell wrote: »

There are several criticisms.

1. It is based on routing codes that have no relevance on the ground. Some are non-contiguous, and many cross county boundaries.

2. The routing codes are over large in some areas and quite small in others.

3. The routing codes are based on a subset of An Post's postal towns. They have about 2,000 of them but have reduced them to about 100.

4. The retention of Dublin post codes was part of the spec but should have been incorporated differently as it has led to too large an area for each single routing code.

5. The random element is just daft.

These criticisms amount, in sum, to "I would have designed it differently".

I have precisely one requirement of a postcode: that it tell me the precise location of the premises to which it refers. Eircode meets that requirement, and as such is working very nicely for me.

6. It appears to be designed to maximise its use for monetizing the database.

This criticism amounts to "I want to be able to access it without paying for it". Which you can, unless you need to access more than 15 per day, at which point it's hard to argue that your requirements don't justify paying for the information.

Although the code is designed as a postcode, it could have been the basis for many other uses, such as navigation, statistical analysis, etc.

It still can. You have convinced yourself that such uses could only have been made of a code if all the relevant information was stored in the code itself. It's actually a lot more efficient to store that information in a database.

MBSnr wrote: »

Dated today.

http://www.irishexaminer.com/ireland/emergency-services-brand-eircode-worse-than-useless-365758.html

However, one senior fire officer said: “As it is, Eircode is worse than useless.”

"Worse than useless" implies that it's actively harmful in some way; that the job of the emergency services has been made more difficult by Eircodes. Given that he doesn't explain how this is the case, I think it's fair to say that he's suffering from the same reflexive criticism problem that we've seen all too much of in this thread.

Sam Russell

I suggested in the origins postcode thread that the design should be based on the Telecom Éireann STD codes. These are quite logical, are not county based but more regional. It also has the advantage that most people are familiar with them. For example, most would know Dublin is 01 and Cork is 021, Limerick 061 and Galway 091.

A county based system as suggested in post #1 has merits as well.

gctest50

oscarBravo wrote: »

...........Which you can, unless you need to access more than 15 per day,...............

It has been increased to 50 searches a day

You have 49 searches left today.

plodder

With the benefit of hindsight, here's how I would have done it, for what it's worth...

1) Forget fitting in with An Post's system. They aren't using it (meaningfully) anyway. There has been little or no benefit from fitting in with it.

2) First one or two characters would be county based on car reg system. D for Dublin etc.

3) Next one or two chars would be a large area within county (postal district in Dublin). So, together the first three characters always identifies a large area within a county.

4) Next two characters identifies a CSO small area.

That's your postcode in 5 characters. Then you would have a unique 3 character
code added to this to identify properties. It could be random or not random. Doesn't matter. The 8 character code would be called a location code for situations when you want to specify an exact location. But, the postcode would not .

A file containing postcodes and their centroid location would be provided for free. The full file of 8 character codes would be the paid for product. Anyone wanting to verify addresses would need this. But, there is a whole swathe of uses for postcodes that don't need this level of detail. And they should (have been) facilitated at no cost.

larchill

Numskulls, the lot of them! Thats the way it should have been done, an easy to use & understand system. When you see the shenanigans revealed on Monday night with the behaviour of some of our councillors its no surprise.:mad:

oscarBravo

plodder wrote: »

With the benefit of hindsight, here's how I would have done it, for what it's worth...

1) Forget fitting in with An Post's system. They aren't using it (meaningfully) anyway. There has been little or no benefit from fitting in with it.

2) First one or two characters would be county based on car reg system. D for Dublin etc.

3) Next one or two chars would be a large area within county (postal district in Dublin). So, together the first three characters always identifies a large area within a county.

4) Next two characters identifies a CSO small area.

That's your postcode in 5 characters. Then you would have a unique 3 character
code added to this to identify properties. It could be random or not random. Doesn't matter. The 8 character code would be called a location code for situations when you want to specify an exact location. But, the postcode would not .

A file containing postcodes and their centroid location would be provided for free. The full file of 8 character codes would be the paid for product. Anyone wanting to verify addresses would need this. But, there is a whole swathe of uses for postcodes that don't need this level of detail. And they should (have been) facilitated at no cost.

Like the OP, it's an interesting design. I can see a couple of flaws straight away: first, it's a postcode as opposed to a geocode, so you'd be subject to the same criticism as Eircode from people who believe postcodes should be able to identify individual telephone poles.

Secondly, small areas are subject to revision from census to census, so a code derived from small areas will either mean premises having to change postcode from time to time, or premises with a postcode that refers to a small area they're not actually in. Neither is ideal.

larchill wrote: »

Numskulls, the lot of them! Thats the way it should have been done, an easy to use & understand system. When you see the shenanigans revealed on Monday night with the behaviour of some of our councillors its no surprise.:mad:

I'm at a loss as to the connection between local councillors and Eircodes.

plodder

oscarBravo wrote: »

Like the OP, it's an interesting design. I can see a couple of flaws straight away: first, it's a postcode as opposed to a geocode, so you'd be subject to the same criticism as Eircode from people who believe postcodes should be able to identify individual telephone poles.

Indeed

Secondly, small areas are subject to revision from census to census, so a code derived from small areas will either mean premises having to change postcode from time to time, or premises with a postcode that refers to a small area they're not actually in. Neither is ideal.

That point was discussed at length in the previous thread. I think I demonstrated that there would be far less change than people think, given that the vast majority of "change" actually means just the creation of new areas, not affecting existing ones.

Sam Russell

There is a lot to recommend in a 5 character postcode that can locate a small area. It would be easy to remember, and would apply to neighbours which is good. The eight character version would be unique to the property but would only be needed when trying to identify individual properties, like address verification by banks, insurance companies, and Government agencies.

Sounds good to me. The 5 character version would obviously be freely available. The eight character database would be pay to use.

oscarBravo

plodder wrote: »

That point was discussed at length in the previous thread. I think I demonstrated that there would be far less change than people think, given that the vast majority of "change" actually means just the creation of new areas, not affecting existing ones.

I asked CSO. Their response was:

It is envisaged that a certain number of Small Areas will need to be modified after the next census.

Now, this doesn't quantify the problem; but given that a key design criterion for any postcode system should be to make changes in postcode extremely rare, it wouldn't make a lot of sense to design a postcode system that guarantees a percentage of codes will change every few years.

plodder

oscarBravo wrote: »

I asked CSO. Their response was:

It is envisaged that a certain number of Small Areas will need to be modified after the next census.

Now, this doesn't quantify the problem; but given that a key design criterion for any postcode system should be to make changes in postcode extremely rare, it wouldn't make a lot of sense to design a postcode system that guarantees a percentage of codes will change every few years.

That doesn't contradict my point though. The issue isn't really to avoid changing small areas. It is to avoid changing postcodes.

As far as the CSO is concerned, creating a new small area does mean "changing" an existing one, in the sense that its boundaries need to be changed to create the new one, but for the most part, the area would originally have been empty space (because that's where new housing gets built typically). So, few if any postcodes need to change.

oscarBravo

plodder wrote: »

That doesn't contradict my point though. The issue isn't really to avoid changing small areas. It is to avoid changing postcodes.

As far as the CSO is concerned, creating a new small area does mean "changing" an existing one, in the sense that its boundaries need to be changed to create the new one, but for the most part, the area would originally have been empty space (because that's where new housing gets built typically). So, few if any postcodes need to change.

From CSO again:

The number of Small Areas likely to be changed will only be determined following the receipt of the census returns next year. In 2011 we amalgamated Small Areas with less than 25 occupied households and in some cases we split Small Areas where there was a greater than expected number of households. I would imagine 95-98% of Small Areas will be unchanged from 2011, but this is a speculated guess.

It's not the only possible criticism, anyway. Another is the example I cited in the other thread, with the Streetview image of the two neighbouring premises, one of which in your system would have a postcode beginning "MO", and the other "SO". This may not be a problem for some applications; it may be for others.

The point is that every imaginable design has advantages and disadvantages, depending on your individual requirements. The idea that one design is universally better than another is disingenuous.

Sam Russell

oscarBravo wrote: »

From CSO again: It's not the only possible criticism, anyway. Another is the example I cited in the other thread, with the Streetview image of the two neighbouring premises, one of which in your system would have a postcode beginning "MO", and the other "SO". This may not be a problem for some applications; it may be for others.

The point is that every imaginable design has advantages and disadvantages, depending on your individual requirements. The idea that one design is universally better than another is disingenuous.

If you design any postcode system, it will have boundaries, and postcodes will change significantly across such boundaries. D04 becomes D06 across a line. Mayo becomes Sligo or Galway across a line. That is the nature of a postcode.

To think otherwise is disingenuous.

oscarBravo

Sam Russell wrote: »

If you design any postcode system, it will have boundaries, and postcodes will change significantly across such boundaries. D04 becomes D06 across a line. Mayo becomes Sligo or Galway across a line. That is the nature of a postcode.

To think otherwise is disingenuous.

I know. Happily, I don't care. I'm not the one arguing about how useful a code would be if it were hierarchical.

The theory is that a hierarchical code lets you (say) group deliveries by similar codes. So you start with the county code, and bundle all the SO deliveries together; separately, you bundle all the MO deliveries together. By which logic, you have one truck delivering to the house that's in Charlestown, and another truck delivering to the petrol station next door in Bellaghy.

The idea that a hierarchical code solves all the problems people are complaining about with Eircodes is the proposition I'm arguing against.

plodder

oscarBravo wrote: »

From CSO again: It's not the only possible criticism, anyway. Another is the example I cited in the other thread, with the Streetview image of the two neighbouring premises, one of which in your system would have a postcode beginning "MO", and the other "SO". This may not be a problem for some applications; it may be for others.

But, that's true with any postcode system including Eircode. Two neighbouring houses in Sligo could have Eircode routing keys of F91 or F56 right at the boundary between the two areas. At least everyone knows their own county.

The point is that every imaginable design has advantages and disadvantages, depending on your individual requirements. The idea that one design is universally better than another is disingenuous.

Usually, there is an optimal design for a specific set of requirements. What happened here is that some of the original requirements were discarded and replaced with others, some of which have never been publicly acknowledged, such as the requirement for random codes to make the system not very useful without licensing the database that makes sense of the codes.

hans aus dtschl

I'll start by saying I've absolutely no skin in this game whatsoever: I'm just an onlooker.

Personally, from an end-user perspective, I'd have preferred the Eircodes to be more easily memorable.

It's fine to say the people of Dublin remember their Eircodes: of course they do: they had some of it already, and it starts with D. The rest of the country has an entirely new string of characters/numbers.

Simple things could have improved usage I think: using County letters as described in previous posts, makes more sense to me. Also if the "Eircode Finder" website allowed a basic (non-property-specific) browse of the postcode grid, to remember that Limerick is "V" and Cork is "T" it would help.

I think that hoping that people will remember a code, most of which nobody around them shares, that none of them are (legally) compelled to use was very optimistic. The idea that people shouldn't inherently know postcode areas to avoid "property-value-by-postcode" probably killed end user take-up a bit.

And as people before me have said, it appears that in the list of priorities, making money from Eircode seems to have ranked higher than end-user take-up. I know many people who would benefit from using a post code - living in the countryside - who don't use it. For those implying that "silly culchies won't change": the changeover to the euro was pretty seamless, the smoking ban came in quickly, etc etc. So while people do resist change, to blame end users for slow take-up isn't the whole story at all IMO.

Eircode will now need to do more to encourage usage. Because a lot of people (myself ashamedly included) have looked at their Eircode, looked it up a few times, found the entire thing a little cumbersome, and disregarded it.

Some criticism of the new system (despite all the valid technical rebuttal in the world) is definitely warranted.

marmurr1916

hans aus dtschl wrote: »

I'll start by saying I've absolutely no skin in this game whatsoever: I'm just an onlooker.

Personally, from an end-user perspective, I'd have preferred the Eircodes to be more easily memorable.

It's fine to say the people of Dublin remember their Eircodes: of course they do: they had some of it already, and it starts with D. The rest of the country has an entirely new string of characters/numbers.

Simple things could have improved usage I think: using County letters as described in previous posts, makes more sense to me. Also if the "Eircode Finder" website allowed a basic (non-property-specific) browse of the postcode grid, to remember that Limerick is "V" and Cork is "T" it would help.

I think that hoping that people will remember a code, most of which nobody around them shares, that none of them are (legally) compelled to use was very optimistic. The idea that people shouldn't inherently know postcode areas to avoid "property-value-by-postcode" probably killed end user take-up a bit.

And as people before me have said, it appears that in the list of priorities, making money from Eircode seems to have ranked higher than end-user take-up. I know many people who would benefit from using a post code - living in the countryside - who don't use it. For those implying that "silly culchies won't change": the changeover to the euro was pretty seamless, the smoking ban came in quickly, etc etc. So while people do resist change, to blame end users for slow take-up isn't the whole story at all IMO.

Eircode will now need to do more to encourage usage. Because a lot of people (myself ashamedly included) have looked at their Eircode, looked it up a few times, found the entire thing a little cumbersome, and disregarded it.

Some criticism of the new system (despite all the valid technical rebuttal in the world) is definitely warranted.

My UK mobile phone number has 11 characters. Apart from the first two (which are 07 - all UK mobile phone numbers start with 07), the following 9 characters are completely randomised. I can remember that number, and I can remember my previous UK mobile phone number.

I also remember my UK postcode which contains seven characters, with the first two being letters that bear no relation to the town or county I live in, and no relation either to the nearest large town.

If people can remember that they live in D04 (or Dublin 4 under the old system), how difficult can it be to remember that they live in V92? Only the last four characters of an Eircode are truly random. For the majority of people, the first three characters of Eircodes are shared with the people around them. If people can remember that they live in the 021 or whatever area code, why can't they remember that they live in the T12 routing code area? Is there something about the first three characters of Eircodes that make them uniquely difficult to recall?


I don't see why the randomisation of the final four characters of an Eircode should make it any more difficult to remember than a mobile phone number with more than four randomised characters.