nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Child Item

Eircode design

PDVerse

Hello folks. If anyone has a question about any aspect of Eircode design I'm happy to answer.

Find more posts tagged with

Quick Links

Comments

TheChizler

sdanseo wrote: »

I still don't understand why we couldn't just have a 4-digit postcode system. Works perfectly for the entire continent of Europe - I work in distribution so use them all daily.

But no, we just had to do our own, misguided, STUPID thing. As usual.

How would that sort the non unique addresses, which was the primary requirement of the design brief?

clewbays

PDVerse wrote: »

There isn't an official Routing Key boundary map. Technically boundaries don't exist. A boundary would imply you could infer the Routing Key of a new build in advance.

I appreciate boundaries would limit the flexibility of An Post which understandably absolutely nobody wants to do or even suggest. So An Post can balance the work load on their delivery offices by assigning routing keys in response to demands. Some addresses on the northside of the Liffey are in Dublin 8. Some routing keys are non-contiguous. This flexibility can make business difficult for other users and it means GIS files cannot be mapped to routing keys as there are no shape files available.

So back to the start why didn't you just forget An Post when designing the routing keys given that they have no interest in them and slowed implementation down for so many years?

PDVerse

The design specification required Routing Keys that were suitable for the USP manual sortation requirements. Your characterisation of An Post isn't one that agrees with my personal experience on the project.

Anyone who licenses the ECAD database can create GIS boundaries if they wish to do so. Gamma's creation has stalled due to a lack of priority, not for any technical reason. I believe one of the other VARS supplies Routing Key boundaries if you license the ECAD data from them.

oscarBravo

sdanseo wrote: »

I still don't understand why we couldn't just have a 4-digit postcode system. Works perfectly for the entire continent of Europe - I use many of them daily for work.

Countries with four-digit postcodes - such as Denmark - have unique addresses. We don't.

clewbays

PDVerse wrote: »

I believe one of the other VARS supplies Routing Key boundaries if you license the ECAD data from them.

I am not familiar with the various VARs. Would it be possible for you to ask them to post an example pdf file showing how the Phoenix Park is allocated between the routing key boundaries?

We are going to end up with more and more little pieces of non-contiguous boundaries which will make the routing keys less and less useful for geographers.

sondagefaux

sdanseo wrote: »

I still don't understand why we couldn't just have a 4-digit postcode system. Works perfectly for the entire continent of Europe - I use many of them daily for work.

But no, we just had to do our own, misguided, badly thought out, privatised thing, and throw taxpayer's money at it to prop it up. As usual.

It's not privatised - it's outsourced, a rather crucial difference. The design, implementation and management of a postcode system would not have been free, no matter what the design used, so your comment about 'taxpayer's money' is pointless. As for four-digit postcodes being used across Europe, that's simply not true. France uses 5-digit postcodes, which aren't as precise as UK postcodes (up to 7 alpha-numeric characters). For example, the Paris postcode 75008 tells you that the address is in Paris (75 is its D partement number) in the 8th Arrondissement. Basically it's saying that the address is in Paris 8, which is equivalent in its lack of precision to the Dublin postal districts, Dublin 8 etc.

Prior to 1972, an address in the eighth arrondissement in Paris, would be written as:

8 rue ChambigesParis 8^e
This number was incorporated into the postal code as:
8 rue Chambiges75008 Paris

https://en.wikipedia.org/wiki/Postal_codes_in_France

Germany also uses 5-digit postcodes:

https://en.wikipedia.org/wiki/List_of_postal_codes_in_Germany

If you can remember your 7-digit mobile phone number I'm sure you can remember your 7-character Eircode (or UK postcode) just as easily.

sondagefaux

sdanseo said:
I still don't understand why we couldn't just have a 4-digit postcode system. Works perfectly for the entire continent of Europe - I use many of them daily for work.

Andorra uses 5-character postcodes, the letters AD followed by three numbers.

Belarus uses 6-digit postcodes.

Bosnia uses 5-digit postcodes.

Russia uses 6-digit postcodes.

Croatia uses 5-digit postcodes.

Estonia uses 5-digit postcodes.

The Faroe Islands use 3-digit postcodes.

Finland uses 5-digit postcodes.

France uses 5-digit postcodes.

Germany uses 5-digit postcodes.

Gibraltar uses a 7-character alpha-numeric postcode.

Greece uses 5-digit postcodes.

Holy See (Vatican) uses a 5-digit postcode.

Italy uses 5-digit postcodes.

Kosovo uses 5-digit postcodes.

Lithuania uses 5-digit postcodes.

Malta uses 7-character alpha-numeric postcodes.

Monaco uses 5-digit postcodes.

Montenegro uses 5-digit postcodes.

Netherlands uses 6-character alpha-numeric postcodes (4 digits, space, 2 letters).

Poland uses 5-digit postcodes.

Portugal uses 7-digit postcodes.

Romania uses 6-digit postcodes.

Russia uses 6-digit postcodes.

San Marino uses 5-digit postcodes.

Slovakia uses 5-digit postcodes.

Serbia's old system used 5-digit postcodes, its new system uses 6-digit postcodes.

Spain uses 5-digit postcodes.

Sweden uses 5-digit postcodes.

Turkey uses 5-digit postcodes.

Ukraine uses 5-digit postcodes.

Source: http://www.grcdi.nl/gsb/world%20postal%20code%20formats.html

PDVerse

Clewbays, allmapdata.com have Routing Key boundaries.

I have yet to see a convincing use case where Routing Key boundaries are a good starting point for any geographical analysis. Small Areas from OSi are available from the Open Data website data.gov.ie

sondagefaux

This site not only shows the formats and numbers of characters/digits used in postcodes but also tells you what level of granularity each postcode has.

[justify][font=Georgia, serif]The level of coverage is shown of the following scale, and is a guideline only. For fuller details refer to the appropriate country chapter:[/font][/justify]
[justify][font=Georgia, serif]4 = All postal codes are unique point-of-delivery codes or cover approximately equal-sized groups of points-of-delivery[/font][/justify]
[justify][font=Georgia, serif]3 = Some postal codes down to point-of-delivery or small groups of points-of-delivery; others to settlement or groups of settlement level[/font][/justify]
[justify][font=Georgia, serif]2 =All postal codes cover a settlement, municipality, post office, postal route or larger area[/font][/justify]
[justify][font=Georgia, serif]1 = A single postal code covers the whole country/territory.[/font][/justify]

Ireland has a system where all postal codes are unique point-of-delivery codes.
Only 11 countries and territories (Argentina (8-character alpha-numeric postcodes), Chile (7-digit postcodes), Guernsey, Jersey, the Isle of Man, the UK (up to 7-character alpha-numeric postcodes), Honduras (6-character alpha-numeric postcodes), Iran (10-digit postcodes) Lebanon (new postcodes are 8-digits), Netherlands (6-character alpha-numeric postcodes) and Serbia (new 9-character alpha-numeric postcodes)) have postcodes that cover approximately equal-sized groups of points-of-delivery.
For anyone interested, the level of detail about postal addresses and the use of sub-national divisions in postal addresses (e.g county names) is very comprehensive:
http://www.grcdi.nl/gsb/united%20kingdom.html

sondagefaux

[justify][font=Georgia, serif]Iran has replaced its 5 digit postal code with a ten digit code, intended to give an individual code to each building in the country. The first digit is a forwarding code, the second a zone, the third a district, the fourth a part of the district, the fifth a segment. The final four digits are an identification code.[/font][/justify]

Iran

sondagefaux

[justify][font=Georgia, serif]Lebanon began introducing postal codes in 2000, and the new system will be rolled out over the whole country in the next years. The new postal code contains four consecutive digits in rural areas (and for post offices and government offices) and eight digits in urban areas in the format:[/font][/justify]
[font=Lucida Console,Courier New,Courier,monospace] 9999[ ]9999
[/font]
[justify][font=Georgia, serif]The first four numbers indicate the postal zone , a sector within a town, or a village. The final four digits indicate the postal code extension , which pinpoint a single building within a postal zone. For areas with an eight-digit postal code, those odd-numbered codes indicate private residences, even-numbered codes indicate businesses. Large users (LVR s - Large Volume Receivers) and post office boxes also have eight-digit postal codes.[/font][/justify]

[justify][font=Georgia, serif]Lebanon[/font][/justify]

sondagefaux

[justify][font=Georgia, serif]The postal codes consist of a block of 5 digits. The code is the same as that used before Slovenia, Croatia, Bosnia-Hercegovina, Macedonia and Montenegro left the federation. Serbian codes begin with 1, 2 or 3.[/font][/justify]

[justify][font=Georgia, serif]Serbia had planned to replace this code with a new, 6-digit, code, but is instead (possibly for a transition period) using both within the address block. Each 6-digit code, called Postal Address Code (Po tanski adresni kod, PAK) covers a small group of buildings or a single company. The first two digits cover a region of one or several municipalities, the next two digits define the delivery zone. The final two indicate the street, part of street, building or company. This code is written on its own at the bottom of the address block.[/font][/justify]

Serbia

sondagefaux

[justify][font=Georgia, serif]Postal codes (C digo de Enderecamento Postal, CEP, introduced in 1992) consists of 8 digits in the format:[/font][/justify]
[font=Lucida Console,Courier New,Courier,monospace] 99999[-]999
[/font]
[justify][font=Georgia, serif]The first digit indicates the region, the second a sub-region, the third a sector, the fourth a sub-sector, the fifth a part of a sub-sector, and the numbers after the hyphen indicate the point of delivery.[/font][/justify]

[justify][font=Georgia, serif]The final three digits range between 000 and 899 for street addresses; 900 to 959 for special codes ; 960 to 969 for promotional postal codes ; 970 to 989 and 999 for postal use; and 990 to 998 for post office boxes.[/font][/justify]

Brazil

sondagefaux

The four posts above give information about postcodes in other countries apart from Ireland where either all buildings (Iran, Brazil) or some buildings in some instances (Lebanon, Serbia) can be individually identified from postcodes.

The UK also has special postcodes that can be used to identify specific buildings or specific departments within organisations, e.g. postcodes used by the DVLA (Driver and Vehicle Licensing Agency) in Swansea.

Apart from Ireland, only two other countries, Iran and Brazil, have postcodes which always allow individual buildings to be identified from postcodes. Ireland seems to be unique in having postcodes which can identify individual units within some commercial buildings (e.g. individual shop units within some shopping centres) or individual apartments within many residential buildings.

plodder

PDVerse wrote: »

Clewbays, allmapdata.com have Routing Key boundaries.

I have yet to see a convincing use case where Routing Key boundaries are a good starting point for any geographical analysis. Small Areas from OSi are available from the Open Data website data.gov.ie

Routing key boundaries might have been useful because they are the one public domain aspect of Eircode. But, because of the enormous disparity in routing key area size, they are of little value.

Most (if not all) of those postcode designs in the posts above appear to be hierarchical to at least three levels. That's what we should have got too. Unique codes, while they do present their own issues, aren't the main problem. Also, not sure that Brazilian postcodes are unique down to the level of being potentially personal data. It seems only public buildings/spaces and large private buildings have their own code.

GJG

Sam Russell wrote: »

This is a separate point. Eircode does not help those without access to the database find a property - whether they have the Eircode or not. Armed with just the non-unique address (and sometimes a unique address) find a property.

Every local road in every townland should be given a name or number, and every house should have a distance based number measured from one end of the road. Result would be end of non-unique addresses.

Sam, I don't mean to be disrespectful, but that's an awful lot of wrong to squeeze into a short post.

Eircode does not help those without access to the database find a property

Actually it does. Rural posties already match names to houses by heart, but this breaks down when the same name repeated. Postmen have described knocking on the door to ask if they are expecting a letter from the VHI or whatever. Now the postman to remember which Jimmy Murphy is the one whose Eircodes ends F7 and which ends 4K.

does not help those without access to the database

That would be people without a smartphone? Is it really such a catastrophe that the postman would pull out his phone and look at the Autoaddress app the first couple of days he has a letter for a new resident? Or better asking around has anyone heard of a new family moving in, as has happened in my memory?

Every local road in every townland should be given a name or number

The British changed the boundaries of Counties Galway, Clare, Mayo, Roscommon, Sligo, Waterford, Kilkenny, Meath and Louth, and others in 1898. Even today, nearly 120 years later, people from certain townlands on the Carlow/Kilkenny border insist on the right to play hurling for their 'old' county and football for their 'new' one. If you are saying that this can be done without decades of controversy and delay, you are either hopelessly naive or trolling.

I am familiar with the machinations that went on in the department when Eamon Ryan tried to introduce postcodes. The opposition was covert, and took the form of pushing impossible versions to derail the project; for example basing postcodes on English-language names to trigger opposition from the gaelgeoir . All the why-didn't-they-just posts are basically just the impossible versions from back then.

and every house should have a distance based number measured from one end of the road

And you would also have billions of valid-but-vacant postcodes, making the system useless for a wide variety of uses such as insurance and social welfare security. The problem with the why-didn't-they-just ideas is that they take account of the pet peeve of the writer, but ignore a bunch of other uses.

Result would be end of non-unique addresses.

Actually, it wouldn't. It's the same problem as with the code promoted by the loon on twitter, either it is not granular enough, and two cottages with doors side-by-side would end up with the same code, or it's too granular and even modest-sized properties have several valid codes. I'm aware that there are bunch of uses where that isn't a problem. Those suggestions would do fine for those uses. But there are other uses where that is a problem.

Lots of people can come up with a solution that suits them. That isn't the aim. The aim is a solution that suits all uses.

sondagefaux

If Sam's proposal went ahead, signs would be required for all the roads with their new names. He also said there would be signposts pointing the way to each townland.

There are over 49,000 townlands in the republic of Ireland.

I've already provided information about the cost of signage.

A figure of 100 million is a reasonable estimate of the cost of providing the required signage in townlands to implement Sam's proposal.

And for what? To solve a problem that Eircodes have already solved.

05eaftqbrs9jlh

Might have been asked already but what is the deal with a load of addresses in Leinster being As Gaeilge?!

sondagefaux

Because every address in the republic has two versions, the Irish langauge version and the English language version.

Carawaystick

sondagefaux wrote: »

Because every address in the republic has two versions, the Irish langauge version and the English language version.

Well every address has an Irish language version anyway, and most have English versions.

antoinolachtnai

GJG wrote: »

Actually it does. Rural posties already match names to houses by heart, but this breaks down when the same name repeated. Postmen have described knocking on the door to ask if they are expecting a letter from the VHI or whatever. Now the postman to remember which Jimmy Murphy is the one whose Eircodes ends F7 and which ends 4K.

VHI has not and will not code non-unique addresses. If you are familiar with the project you know this perfectly well.

I am familiar with the machinations that went on in the department when Eamon Ryan tried to introduce postcodes. The opposition was covert, and took the form of pushing impossible versions to derail the project; for example basing postcodes on English-language names to trigger opposition from the gaelgeoir . All the why-didn't-they-just posts are basically just the impossible versions from back then.

There is a clear and obvious way to avoid this problem.

And you would also have billions of valid-but-vacant postcodes, making the system useless for a wide variety of uses such as insurance and social welfare security.

The present system similarly has 'billions' of valid-but-vacant postcodes. It does not make the system useless.

[/quote]
Lots of people can come up with a solution that suits them. That isn't the aim. The aim is a solution that suits all uses.[/QUOTE]

And as a result we have a system that is designed to solve a commercial and intellectual property problem, but doesn't really suit any logistical use. (It can be shoehorned to a logistical use, but it doesn't actually suit it.)

GJG

I am, and I have no idea what you mean by this, could you supply a source?

There is a clear and obvious way to avoid this problem.

Could you share it with us?

The present system similarly has 'billions' of valid-but-vacant postcodes. It does not make the system useless.

No, there are 2.2m codes in the system in total. There are billions of combinations that match the Eircode format, but if I use invalid ones of them to, for example, disguise the fact that I'm insuring my house multiple times to commit fraud, the invalid ones are instantly evident.
By contrast, Loc8 has at least six valid codes that point to my small house, including one that is shared by my and my neighbour's front door. Uniquely identifying each dwelling is an obviously useful tool in deterring and detecting fraud.

Lots of people can come up with a solution that suits them. That isn't the aim. The aim is a solution that suits all uses.

I couldn't agree more. Can point out a viable design change within Eircode's remit that can a) improve its performance in some way b) without prejudicing other aspects of its use?

antoinolachtnai

GJG wrote: »

I am, and I have no idea what you mean by this, could you supply a source?

Could you share it with us?

The use of numbers?

No, there are 2.2m codes in the system in total. There are billions of combinations that match the Eircode format, but if I use invalid ones of them to, for example, disguise the fact that I'm insuring my house multiple times to commit fraud, the invalid ones are instantly evident.
By contrast, Loc8 has at least six valid codes that point to my small house, including one that is shared by my and my neighbour's front door. Uniquely identifying each dwelling is an obviously useful tool in deterring and detecting fraud.

I did not realise we were comparing eircode to loc8. I thought we were moved on to the case that eircode was better than any other possible code.

There is an obvious solution to that too. Have a database of valid codes.

I couldn't agree more. Can point out a viable design change within Eircode's remit that can a) improve its performance in some way b) without prejudicing other aspects of its use?

You quoted something someone else said as being something I said. I did not say that. I said that the system as designed is fairly useless. It is not completely useless, it can be shoehorned to various tasks, but it is fairly useless.

The design changes I can propose to eircode are very limited, because it has already been rolled out. The metaphorical concrete has set. There is nothing that can really be done about it now.

But one change I would propose is to make the code open and stop the nonsense of charging for access to the database, in particular the database linking eircodes to addresses and statistical small areas. This could be done to a greater or lesser extent, but it would be a great help to getting the code into use and reaping more benefits.

oscarBravo

antoinolachtnai wrote: »

You quoted something someone else said as being something I said. I did not say that. I said that the system as designed is fairly useless. It is not completely useless, it can be shoehorned to various tasks, but it is fairly useless.

You, and every other Eircode critic, could really do with learning the difference between "not what I would have designed" and "useless".

I have found Eircode useful for several different functions in my business. I haven't had to "shoehorn" it to get utility from it; I've used it as it was designed.

If something is useful, it can't be useless. It seems illogical to even describe it as "fairly useless", unless you're using it as shorthand for "useless for a carefully-selected subset of use cases that I've cherrypicked", in which case it's more disingenuous than illogical.

But one change I would propose is to make the code open and stop the nonsense of charging for access to the database, in particular the database linking eircodes to addresses and statistical small areas. This could be done to a greater or lesser extent, but it would be a great help to getting the code into use and reaping more benefits.

It's kind of ironic that, at the same time as Eircode is being criticised for how much it cost to implement, it's also being criticised for including a mechanism to recoup much of that cost.

plodder

antoinolachtnai wrote: »

But one change I would propose is to make the code open and stop the nonsense of charging for access to the database, in particular the database linking eircodes to addresses and statistical small areas. This could be done to a greater or lesser extent, but it would be a great help to getting the code into use and reaping more benefits.

I agree to the extent that -

a) it's only fair that Eircode should provide free access to the same kind of information that you can get for free from other countries' postcodes.

b) they could be providing multiple different products, at different price points for different purposes and maybe would end up earning more out of it than the model they have now.

It does appear that this idea (or a version of it) was floated, but may have been shot down, since it needed a renegotiation of the geodirectory license, which wouldn't bode well for Eircode adapting to other changes in the market in years to come. This is all speculation of course, because nobody outside of the project really knows what goes on inside.

antoinolachtnai

Calling me disingenuous is a bit strong.

The code is basically useless for delivering post or other services. Ask any postman or delivery man.

It doesn't help you sort the mail because it does not identify localities (beyond the very large routing code areas). This is just a fact. (You can of course pay a fee and do a lookup to get useful information about the locality, but then you are really using the database, not the code.)

I am happy that you have found uses for the eircode. I don't know what your use case is but I am inclined to think that you have found uses for the address database and unique house identifier, rather than eircode as a code that is useful for actually delivering things.

That is why I am saying that the eircode is, in and of itself fairly useless. The database of addresses is useful for sure. But the database was useful even before eircode was introduced (and for sure, eircode has been an enhancement to this database).

I am not criticising eircode for having a revenue model. I would on the other hand be critical of the fact that it is not a very good revenue model.The revenues are going to be very small, because the code has such little uptake.

(If you have information that the revenues are turning out to be substantial, I would be interested in hearing more.)

oscarBravo

antoinolachtnai wrote: »

The code is basically useless for delivering post or other services. Ask any postman or delivery man.

My business delivers services using Eircodes. Your claim is factually inaccurate.

It doesn't help you sort the mail because it does not identify localities (beyond the very large routing code areas). This is just a fact. (You can of course pay a fee and do a lookup to get useful information about the locality, but then you are really using the database, not the code.)

This is why I think it's far from strong to describe your argument as disingenuous.

I do pay a fee, and I do get useful information from the database. To describe that as "using the database, not the code" is the most ludicrous goalpost shift I've seen on this topic to date - and that's saying something.

Pray explain to me: how do I look up the database, if not by using the code?

I am happy that you have found uses for the eircode. I don't know what your use case is but I am inclined to think that you have found uses for the address database and unique house identifier, rather than eircode as a code that is useful for actually delivering things.

My primary use case is delivering services, so you're wrong. An eircode allows me to identify a customer's premises with 100% accuracy and without wasting time trying to decipher directions.

The database of addresses is useful for sure. But the database was useful even before eircode was introduced (and for sure, eircode has been an enhancement to this database).

The database of addresses was useless for my purposes without Eircodes. Having a key to that database in a format that people can easily memorise is invaluable.

Granted, not everyone has memorised theirs - yet. Perhaps they'd be quicker to do so if they didn't have a constant stream of uninformed opinion telling them how useless the codes are.

I am not criticising eircode for having a revenue model. I would on the other hand be critical of the fact that it is not a very good revenue model.The revenues are going to be very small, because the code has such little uptake.

(If you have information that the revenues are turning out to be substantial, I would be interested in hearing more.)

I don't know what the revenue streams are like. It seems that you don't either - but that doesn't seem to be stopping you from making claims about them.

antoinolachtnai

If you don't agree there is a distinction between the code and the database then I can't really argue with you.

I can't understand how a code which no one knows and almost no one uses has made the address database useful for you. I cannot see how a code which identified localities wouldn't have been equally or more useful for your application. Still I am glad you are getting value out of eircode.

The fees charged are very low in comparison to the expenditure and the business uptake is pretty small. There is not going to be much revenue in it.

I am sorry you have resorted to personal attacks instead of discussing the issues. Maybe you just don't understand the words you are using but I reject the personal remarks you have made in the strongest terms.

oscarBravo

antoinolachtnai wrote: »

If you don't agree there is a distinction between the code and the database then I can't really argue with you.

I understand the difference. The problem is that you're trying to argue that a system is useless by taking one part of it in isolation and claiming that that one part isn't particularly useful if you insist on refusing to use the rest of the system.

It's a bit like arguing that telephone numbers are useless without access to a phone network. Well, duh. Phone numbers aren't intended as a means of contacting someone without using the network to which they're connected. Eircodes aren't designed to be useful without the relevant databases.

So, no: Eircodes aren't useless when they're used as designed. They could well be useless if you insist on trying to use them in ways they weren't designed to work, but that tends to be a truism about almost anything, and isn't as insightful as you seem to think it is.

I can't understand how a code which no one knows and almost no one uses has made the address database useful for you.

The problem is that you're operating exclusively on the basis of your assumptions, and it's solely on the basis of those assumptions that you've declared Eircodes useless.

Sure: Eircodes would be useless to me if nobody knew or used their code. And yet, they're useful to me. There's a logical conclusion in there, if you bother to look for it.

I cannot see how a code which identified localities wouldn't have been equally or more useful for your application.

I don't need to know my customers' localities; I need to know their location. Seriously: how the hell is it remotely useful for the delivery of a service to have a vague idea of roughly where someone is? The value to me of Eircodes is their precision.

...the business uptake is pretty small.

How small?

I am sorry you have resorted to personal attacks instead of discussing the issues. Maybe you just don't understand the words you are using but I reject the personal remarks you have made in the strongest terms.

Personal attacks? You're offended because I've pointed out flaws in your reasoning?

antoinolachtnai

Your choice of language was personal and insulting and once again I reject the insinuation.