National Postcodes to be introduced

GoCoder

recedite wrote: »

I think the high maintenance costs were due to a perceived need for the database to be constantly updated, whereas a system that is self updating/verified by the users should not be costly, once set up.

More observations on Loc8 V GO Code;
If you use a less precise area code, using only 5 digits, then the map on the GO Code website merely pinpoints a less accurate position.

The map on the Loc8 website indicates a more useful general area square of 100m X 100m. Using only 3 digits shows a square of 3.5 X 3.5 Km.

The eighth (extra) digit on the Loc8 code is a "checker" number. So no wasted journeys for courier drivers due to a sloppy transcription error.

If you key in the GOCode; END WRLD you end up somewhere on the road between Navan and Mullingar. Dublin based company?

Recedite -
The Minister's comments were based on the Govt postcode design of ABC 123 which requires a database to be maintained.

Self-updating systems are not that reliable on their own. Look at Google address database.

Using 5 characters from a GO Code will actually give you a defined area, of which that 5 character GO Code is the centre-point. I can't recall offhand what the size of the area is but can find out.

ENDWRLD would put you in the middle of a field, near the N52. Maybe the owner might like it. Good spot though.

antoinolachtnai

The Go Code as proposed still requires a database of addresses which needs to be maintained, no?

liam84

antoinolachtnai wrote: »

The Go Code as proposed still requires a database of addresses which needs to be maintained, no?

Eh no. Or should you say yes to a negative question?

Yes. I hope.

But no as well, just in case I'm wrong)

baalthor

MYOB wrote: »

They added an extra 4 digits a long, long time ago.

A US poster (Overheal I think) posted on this a few months ago: basically no-one in the US uses ZIP+4

@blackwarrior: other countries have simpler post codes (when compared with Loc8 etc) because their houses and businesses have unique addresses.
We in Ireland have 1. a very large % of houses in rural areas and 2. these houses do not have house numbers and street/road names to identify them.

One of the big advantages of a proper address system is that you know when you've arrived at the right house because the house number and street/road name are there for you to see.

With the location codes being proposed, in a rural area you don't have this certainty - all you know is that you've arrived at a house (hopefully) but how can you be sure it's the house ?

A house number uniquely identifies a property while a location code uniquely identifies a square or polygon in or near the property. So the same property may end up with several different codes depending on how accurately the user identified its location on an on-line map.

recedite

baalthor wrote: »

With the location codes being proposed, in a rural area you don't have this certainty - all you know is that you've arrived at a house (hopefully) but how can you be sure it's the house ?

A house number uniquely identifies a property while a location code uniquely identifies a square or polygon in or near the property. So the same property may end up with several different codes depending on how accurately the user identified its location on an on-line map.

If your sat nav has brought you to a defined point on someone's driveway, and the dog is barking at you and the owner is looking out the window.....then take it that you have arrived!
Seriously though, to prevent some clown uploading the wrong coordinates, a verification system should be in place. Couriers and the like could rate the code as "trusted" after a successful usage. In that way the database maintains itself. Obviously there would be some commercial gain to the company controlling access to the "list" of codes/adresses.

recedite

antoinolachtnai wrote: »

The Go Code as proposed still requires a database of addresses which needs to be maintained, no?

In the early stages especially, not many people have got their code by pinpointing their location on a map. So it uses the An Post Geodirectory to translate a typed address into a map coordinate. Then converts the coordinate into a GoCode. You can do the first part yourself here;
http://www.geodirectory.ie/GeodirectoryMap.aspx
This database is owned by an Post; they can charge commercial customers for using it. Its a good way of initially getting coordinates and therefore codes for most buildings in the country, although it won't work so well in rural areas.

GoCoder

recedite wrote: »

If your sat nav has brought you to a defined point on someone's driveway, and the dog is barking at you and the owner is looking out the window.....then take it that you have arrived!
Seriously though, to prevent some clown uploading the wrong coordinates, a verification system should be in place. Couriers and the like could rate the code as "trusted" after a successful usage. In that way the database maintains itself. Obviously there would be some commercial gain to the company controlling access to the "list" of codes/adresses.

Good point, Recedite. We'll include that function on the search facility on our new site. Thanks for the feedback.

The National Address File on the GO Code site just has addresses and GO Codes attached to them. If people use the Map Dropper to identify a location then the system automatically converts that point into a GO Code for the user. People can change the GO Code we currently have assigned to a building/location by simply mailing us. Some have already done that to include entrances, particularly if the building is far away from mapped road. We'll be adding a lot of POIs on the next update as well and commercial names. The GeoDirectory link you provided doesn't seem to have coordinates, only addresses, so not much use, and it's copyrighted. Perhaps you meant the Ordnance Survey service - www.osi.ie - we're looking at using the Map Genie service they have launched commercially.

If users mail us geo-coordinate details, (not copied from OSI site) we'll happily convert them for people if they want to do it that way. Or maybe we'll just put a conversion tool on the site, so that people can do it themselves and add in the location details. Would that be useful?

Alternatively, if people have got a mobile phone with GPS, they can use that, or see if their phone can use our mobile GO Code app. The iPhone app store has them available for free, and there's new versions on their way shortly for Blackberry, Android, etc.

Sponge Bob

Google maps is less than useless IMO, Map Genie is bang up to date.

GoCoder

Sponge Bob wrote: »

Google maps is less than useless IMO, Map Genie is bang up to date.

Not sure about the useless description, but Map Genie looks good, if a bit jerky in its transitions - confusing for user. Not great on all road/street names either.

Sponge Bob

Google maps does not even show the road I live on

antoinolachtnai

recedite wrote: »

In the early stages especially, not many people have got their code by pinpointing their location on a map. So it uses the An Post Geodirectory to translate a typed address into a map coordinate. Then converts the coordinate into a GoCode. You can do the first part yourself here;
http://www.geodirectory.ie/GeodirectoryMap.aspx
This database is owned by an Post; they can charge commercial customers for using it. Its a good way of initially getting coordinates and therefore codes for most buildings in the country, although it won't work so well in rural areas.

The GoCode is a derivative database of the Geodirectory. It doesn't stand alone. Geodirectory (or another address database) is essential to make GoCode work.

Realistically, if GoCode succeeds in its ambition and start selling its address database including a commonly used postcode, Geodirectory's business will be destroyed and it will be shut down. The Golden Goose will be dead in the water and GoCode will have to cover the cost of maintaining the database going forward, and resolving the many serious problems with the Geodirectory database (which is still a very good database in many ways).

I am not saying that there is anything necessarily wrong with this (although Geodirectory's promoters would have obvious and legitimate misgivings). But the reality is that if you want to have a postcode which pinpoints individual houses and that is reasonably useful and accurate, you will have to have a database of buildings. There is no 'cheap' way to do this job if quality is important.

The same goes for the mapping. You absolutely need a high quality set of comprehensive and pinpoint-accurate maps to be able to sort and deliver on the basis of the PONC/Loc8 and Gocodes. Google maps are not these maps (nor do they advertise themselves as such).

You could have a much simpler system by simply applying 5-digit numbers to all the Electoral Divisions and/or Census Small Areas. This could stand separately from an address database and the address database would not be essential to use it, although it would be a useful added extra. It would bring you down to the nearest 60 or so houses. If you wanted a finer-grained code at a later date, you could just add an extra two digits for the road segment, and a further two digits to identify the house (obviously at this stage, an address database would be essential).

baalthor

recedite wrote: »

If your sat nav has brought you to a defined point on someone's driveway, and the dog is barking at you and the owner is looking out the window.....then take it that you have arrived!

Most houses are un-occupied during work hours (apart from the hundreds of thousands of empties). Not sure about the geo-positional abilities of the average canine although according to some people they know everything

GoCoder

antoinolachtnai wrote: »

You could have a much simpler system by simply applying 5-digit numbers to all the Electoral Divisions and/or Census Small Areas. This could stand separately from an address database and the address database would not be essential to use it, although it would be a useful added extra. It would bring you down to the nearest 60 or so houses. If you wanted a finer-grained code at a later date, you could just add an extra two digits for the road segment, and a further two digits to identify the house (obviously at this stage, an address database would be essential).

That's an interesting idea. Have you developed a system for it?

The Govt is bringing in its own postcode design for sorting and delivery of mail. The OSI Small Areas are likely to be linked to that postcode system that the Govt is bringing in - ABC 123. And GeoDirectory will be used as part of that, according to all the experts. In theory, GeoDirectory could have the national postcode and GO Codes attached to its database and people could use that instead. Is that likely to happen? Selling our database, as you put it, does not make sense. For GO Code to work, it needs codes assigned to locations, of which buildings are an important subset. It works dynamically, so people can add locations to it at will, so of course it needs to be maintained and be added to.

baalthor

baalthor wrote: »

Most houses are un-occupied during work hours (apart from the hundreds of thousands of empties). Not sure about the geo-positional abilities of the average canine although according to some people they know everything

At least with a barking dog, you can be fairly sure the house is lived in, unless the dog's squatting. :P

antoinolachtnai

Gocoder: you have made a number of speculative statements in your second paragraph there and I can't really answer your question as to whether they are true, although I suspect most of them are not.

My point, to use your language, is that whilst almost any other code might be 'linked' to Geodirectory, GoCode is almost wholly a derivative of GeoDirectory. There is not necessarily anything wrong with that and it does have advantages, although it does raise important technical, legal and commercial issues. (and I took it from your previous comments that you had considered and were planning to offer commercial database services, which is what I meant by 'selling' the database).

seawolf145

dannym08 wrote: »

From the RTÉ Website



About time but 2011 seems like a long way off

could they not jus use an existing one like PONC or similar

Welcome to the modern world of 2011 IRELAND

Sponge Bob

antoinolachtnai wrote: »

You could have a much simpler system by simply applying 5-digit numbers to all the Electoral Divisions and/or Census Small Areas. This could stand separately from an address database and the address database would not be essential to use it, although it would be a useful added extra. It would bring you down to the nearest 60 or so houses. If you wanted a finer-grained code at a later date, you could just add an extra two digits for the road segment, and a further two digits to identify the house (obviously at this stage, an address database would be essential).

As it happens there is a 5 digit numbering system for every electoral district ( all 3000 odd of them) run by the CSO/An Post .

There are 50,000 small areas ( SAPS) corresponding to Townlands.

Been there for years which is of course why those two muppets Dempsey and Ryan had a consultation each on postcodes.

antoinolachtnai

Here are some maps that show how a system based on the established divisions (province, county and electoral divisions) might work.

As you can see, some work is required to keep the numbering coherent and meaningful. This is obviously not final, and there are a few variations possible. It would require some public consultation to finalise a design.

This paper is based on electoral divisions, but pretty much the same formula will work for Small Areas.

This system would be inexpensive to establish and would not require an address database at the outset. It would fulfill 90 percent of the requirements of a postcode.

It could be extended to identify individual by numbering the streets and roads within divisions or small areas. This would fulfill almost all the requirements of a postcode. A further two digits would be enough to identify every single property/delivery point. (Obviously at this point an address database would be essential.)

Based on the ergonomics literature and international best practice, numbers and letters are used to maximise accuracy and recall.

GoCoder

antoinolachtnai wrote: »

Here are some maps that show how a system based on the established divisions (province, county and electoral divisions) might work.

As you can see, some work is required to keep the numbering coherent and meaningful. This is obviously not final, and there are a few variations possible. It would require some public consultation to finalise a design.

This paper is based on electoral divisions, but pretty much the same formula will work for Small Areas.

This system would be inexpensive to establish and would not require an address database at the outset. It would fulfill 90 percent of the requirements of a postcode.

It could be extended to identify individual by numbering the streets and roads within divisions or small areas. This would fulfill almost all the requirements of a postcode. A further two digits would be enough to identify every single property/delivery point. (Obviously at this point an address database would be essential.)

Based on the ergonomics literature and international best practice, numbers and letters are used to maximise accuracy and recall.

If I understood your previous post, you don't think it's true that the Govt is going to run with the ABC 123 design. Why don't you submit this for the postcode tender? From your description, it sounds like it would be much more suited for the purpose.

As you say, it wouldn't need a database initially. And it could be linked to GeoDirectory when you get around to identifying single properties.

Presumably the attaching co-ordinates could then be linked to any location code systems that are still around. And we wouldn't need to keep building a National Address File - just focus on non-address locations.

How long would it take to create, allowing for your public consultation period?

antoinolachtnai

Gocoder,

I don't see how there can be a practical ABC 123 code. There are just too many places which begin with the same letters. ( According to the latest census results, there are around 300 towns with a population greater than 1000. Of these, 3 names begin with ‘ATH’ (the example that was used in the publicity). Around 15 begin with ‘BAL’. 9 begin with ‘CAS’, 7 begin with ‘DUN’. 13 begin with ‘KIL’ and a further ‘9’ town names begin with ‘RAT’. A large proportion end with an ‘N’, ‘S’, ‘D’ ‘L’ or silent ‘E’. As a result, a code based on finding a distinctive, unique, memorable abbreviation for each of the postal towns in Ireland is likely to produce more confusion than clarity) The government will have to consider this.

I don't think the scheme they envisage is going to be coincident with small areas and ED's either, because their design concept is based around 'block faces' and the An Post delivery network. These are quite unlike the ED and SA concepts. Again, they will have to consider whether this is the best architecture before they embark on detailed design.

It could be linked to Geodirectory from the very outset, in the sense that every address in Geodirectory is linked to an electoral division.

Certainly, you would be able to move forward linking to Geodirectory as you suggest, at least technically. But there are important commercial issues to consider. Most importantly, who is going to pay to keep Geodirectory on the road if this happens? If there were a building-level postcode, most of the Geodirectory revenue would evaporate overnight, but a national database of buildings and addresses would still be required.

What you say about concentrating on providing building-specific postcodes in places that don't have unique addresses makes very good sense. It would be nice to have a code for every single house, but if a house already has a unique address, it isn't really a priority.

Timing-wise, including public consultations, and to allow plenty of time, you could be ready to go in about 5 months for an ED-based scheme, about 9 months for a small-area based scheme.

If you decided you needed it, you could number the road segments progessively over a five-year period. This would be an add-on. You could launch a perfectly good postcode without this level of granularity.

Sponge Bob

A purely numeric code is equally applicable to Irish or English names. Language agnostic if you will.

If one lives in an Irish speaking area beginning with A or B one will not adopt some technogobble with a start code of QZZ or YJK which are not even letters in the Irish langauge.

One will ask for a translation into Irish...and one even has a constitutional right to do so

GoCoder

Sponge Bob wrote: »

A purely numeric code is equally applicable to Irish or English names. Language agnostic if you will.

If one lives in an Irish speaking area beginning with A or B one will not adopt some technogobble with a start code of QZZ or YJK which are not even letters in the Irish langauge.

One will ask for a translation into Irish...and one even has a constitutional right to do so

You're safe enough, I don't think we have any.:)

GoCoder

antoinolachtnai wrote: »

Gocoder: you have made a number of speculative statements in your second paragraph there and I can't really answer your question as to whether they are true, although I suspect most of them are not.

My point, to use your language, is that whilst almost any other code might be 'linked' to Geodirectory, GoCode is almost wholly a derivative of GeoDirectory. There is not necessarily anything wrong with that and it does have advantages, although it does raise important technical, legal and commercial issues. (and I took it from your previous comments that you had considered and were planning to offer commercial database services, which is what I meant by 'selling' the database).

No - we're not going to do that. We have addresses and GO Codes only. GeoDirectory has a lot more info than that. Adding GO Codes to addresses is what we could do. People give us co-ordinates, we can bulk convert. Most places have GeoDirectory already anyway so the only thing we can provide is a GO Code for use internally.

antoinolachtnai wrote: »

Certainly, you would be able to move forward linking to Geodirectory as you suggest, at least technically. But there are important commercial issues to consider. Most importantly, who is going to pay to keep Geodirectory on the road if this happens? If there were a building-level postcode, most of the Geodirectory revenue would evaporate overnight, but a national database of buildings and addresses would still be required.

What you say about concentrating on providing building-specific postcodes in places that don't have unique addresses makes very good sense. It would be nice to have a code for every single house, but if a house already has a unique address, it isn't really a priority.

People would pay GeoDirectory to keep it updated. Or the postcode management company (maybe that's you with your design??) would. They managed it quite successfully in the UK for a number of years before the recent change in April this year.

If there's a building-level database available with GO Codes, then what I meant by non-address elements is things that don't have an address - POIs if you like.

recedite

antoinolachtnai wrote: »

Realistically, if GoCode succeeds in its ambition and start selling its address database including a commonly used postcode, Geodirectory's business will be destroyed and it will be shut down. The Golden Goose will be dead in the water and GoCode will have to cover the cost of maintaining the database going forward,The same goes for the mapping. You absolutely need a high quality set of comprehensive and pinpoint-accurate maps to be able to sort and deliver on the basis of the PONC/Loc8 and Gocodes. Google maps are not these maps (nor do they advertise themselves as such).

IMO we are heading towards a more "open source" model where the public/small delivery firms update the database themselves in the future. The days of the An Post monopoly are coming to an end, although they will continue to charge for commercial use of their database as long as possible.
Deregulation/liberalisation of the postal system is coming soon, and they don't particularly want to facilitate competitors by contributing to the formation of a rational openly available postcode.
It has been pointed out that the OSI map genie has the best maps (at a price). The An Post Geodirectory was originally jointly set up jointly with OSI, yet it now uses Microsoft Bing mapping. So it appears to be every man for himself out there now.
40% of all addresses in GeoDirectory are non unique and therefore only the owners can tell which is which by clicking on a map. Also, lets not forget that a buildings database is only part of it. For recreational use etc we want a sytem of codes to easily identify and navigate to any location.

antoinolachtnai wrote: »

Here are some maps that show how a system based on the established divisions (province, county and electoral divisions) might work....
As you can see, some work is required to keep the numbering coherent and meaningful.

From your paper; "the second digit is assigned, beginning with 1, to the individual counties after ordering them alphabetically according to the Irish name. For Ulster, the six counties which are part of the United Kingdom have been inserted into the sequence"

Er... I think using prefixes "as gaeilge" could be politically explosive, especially in N Ireland. You might consider investing in one of those grenade proof cages for your offices, like the old RUC stations, if you win the tender.
Anyway, it is more important to have a regular grid of squares or polygons in which the adjacent cells are labelled in a related sequence, not in alphabetical order. The electoral divisions are also too irregular in shape.
I am forming the opinion now that Loc8 has already examined and solved all these issues and is in fact the best system for this country to adopt.

antoinolachtnai wrote: »

who is going to pay to keep Geodirectory on the road if this happens? If there were a building-level postcode, most of the Geodirectory revenue would evaporate overnight, but a national database of buildings and addresses would still be required.

Two words; Property Tax

GoCoder

recedite wrote: »

I am forming the opinion now that Loc8 has already examined and solved all these issues and is in fact the best system for this country to adopt.

Have you a particular expertise in the area? I thought you said the best system was to use the National Grid coordinates.

antoinolachtnai

recedite wrote: »

It has been pointed out that the OSI map genie has the best maps (at a price). The An Post Geodirectory was originally jointly set up jointly with OSI, yet it now uses Microsoft Bing mapping. So it appears to be every man for himself out there now.

This is not a correct statement.

From your paper; "the second digit is assigned, beginning with 1, to the individual counties after ordering them alphabetically according to the Irish name. For Ulster, the six counties which are part of the United Kingdom have been inserted into the sequence"

Er... I think using prefixes "as gaeilge" could be politically explosive, especially in N Ireland.

So use some other order. It doesn't make much difference.

Anyway, it is more important to have a regular grid of squares or polygons in which the adjacent cells are labelled in a related sequence, not in alphabetical order.

The natural and man-made geographic features of the landscape are not square-shaped, and this is obviously the reason why electoral divisions are not square-shaped.

recedite

GoCoder wrote: »

I thought you said the best system was to use the National Grid coordinates.

I did, but an open minded person is always willing to change their opinion as more information and knowledge is obtained.

IMO the folowing are the basic requirements, and they are all fulfilled as well as is possible by Loc8;
1. Be as easy to remember as a phone number.
2. Be uploadable and usable ASAP by the person at the address.
3. Be trusted by the person delivering to the address, be compatible with satnav, and have a checker built into the code which alerts to any error in transcription.
4. Be able to pinpoint any location, not just a building address.
5. Not cost the taxpayer millions to set up and maintain.
6. Be subdivided by a grid into cells which are labelled sequentially to adjacent cells (this is a tricky one). Multiple addresses in the same and neighbouring areas can then be bundled together for efficient delivery of services.
7. Using less digits gives larger cells. Therefore a person can choose to disclose only the rough area they live in, instead of their specific location.
8. Not cause offence to people by using a prefix based on the version of a placename they don't like.


Neither the original government proposed version nor the Go Code (nor the Nat Grid) would seem to satisfy all the requirements.

mackerski

recedite wrote: »

IMO the folowing are the basic requirements, and they are all fulfilled as well as is possible by Loc8;

I missed the requirement that it would be usable by providers of mapping and location-aware tools without either:

1. Charge
2. Reverse-engineering
3. Crowd-sourcing to build an open database from scratch

recedite

GoCoder wrote: »

Have you a particular expertise in the area? .

Loc8 Codes are Grid Coordinates - just in a form which people with no expertise can use without making a mistake and without being mistaken for something else

What are GoCodes?

Is the GoCode: 2222222 a phone number, a grid reference or something a ballet dancer wears?

antoinolachtnai wrote: »

This is not a correct statement.

Which bit do you object to? Please feel free to correct it.:)

GoCoder

mackerski wrote: »

I missed the requirement that it would be usable by providers of mapping and location-aware tools without either:

1. Charge
2. Reverse-engineering
3. Crowd-sourcing to build an open database from scratch

Do I understand from this Mackerski that you would want GO Code to be usable by providers of mapping and location-aware tools and that it would be provided to them free of charge, and you'd want a free database as well?

Or could you clarify?