National Postcodes to be introduced

SpaceTime

If there's a big fee involved it won't be used. Eb Irish market's too small and has functioned without postcodes for decades. I can't see mapping companies being willing to pay that much.

However,Google maps now contains huge amounts or data including house names wherever they got those from.

Schadenfreudia

Sorry for going a teeny weeny bit off topic...but this thread stared in September 2009 and now has over 2,500 comments.

Is this any sort of boards.ie record?

Schadenfreudia

Schadenfreudia wrote: »

Sorry for going a teeny weeny bit off topic...but this thread stared in September 2009 and now has over 2,500 comments.

Is this any sort of boards.ie record?

No, the record is about 10,000 posts. then the thread is killed and a new one started.
Long threads kill the servers so the Admins trim them.

Impetus

a65b2cd wrote: »

Can you tell us more about this average pensioner Impetus - you don't seem to have much regard for them judging by your ageist remarks ...

I am not being "ageist" in this remark. Quite the contrary.

No more than a statement about making public premises accessible to all might be interpreted as being anti people who have to use wheelchairs etc.

I am simply pointing out the stupidity of making life difficult for people, needlessly, when modern computer technology provides the required solution (ie numbering each building) combined with a short, easy to remember postcocde that is used by all in a locality. A solution that works in almost all other European countries, who don't seem to have the same urge, for whatever reason (???) to re-invent the wheel.

Impetus

Deleted User wrote: »

No, the record is about 10,000 posts. then the thread is killed and a new one started.
Long threads kill the servers so the Admins trim them.

It makes one think that the powers that be ignore the so called "Eircode" issue at their peril.

Impetus

I came across this real-world paper on research into the accuracy of machine recognition of all number postcodes - ie those derived from 0 1 2 3 4 5 6 7 8 9, with particular emphasis on machine recognition of hand-written codes.

Using two classification systems, the system can recognise 99.5% of handwritten postcodes, and can perform substitutions of mis-read codes for the other 0.5% of addresses.

Some people on this thread have remarked the absence of a check digit. In real world applications, the rest of the address provides a massive "check-digit" in terms of matching accuracy.

The paper points out the machine complications particularly when dealing with larger than C5 envelopes ("flats") in finding the AoI (area of interest - ie the destination address) in terms of machine recognition, word and character segmentation, and eventual address verification.

Most postal administrations have determined that the word and character segmentation issues can be best dealt with by putting an all numeric postcode on the last line, before the name of the town. In this way, the postcode will always be on the bottom left corner of the address block.

The research was performed by Marcus Pfister of Siemens AG (who provided much of the equipment used by An Post and many other European administrations), as well as Sven Behnke (also Siemens) and Raul Rojas, professor of computing science at Freie University of Berlin.

http://page.mi.fu-berlin.de/rojas/pub/zip_codes_2000.pdf

Machine recognition of computer printed addresses is somewhat less challenging - however a large volume of post and other documents in Ireland are hand written.

The structure of the proposed "Eircode" is a million km away from beieng either machine or human friendly.

Britain has had a large problem recognizing a high percentage of its alpha numeric postcodes - in terms of finding the area of interest and where exactly does the code start and finish and mix-ups between numeric and alpha characters or total failure to read some alpha characters. The people behind Eircode seem determined to follow the same path.

a65b2cd

Interesting article - scanning technology has improved greatly since the late 1990s when that study was done.

The use of a reduced character set for the Eircode (around 25 instead of 36 letters and numbers) will make it very machine and human friendly.

Is the reference to a large volume of hand-written addresses from the 1990s as well - I suppose it is mainly Christmas cards?

SpaceTime

The majority of letters I receive are printed.

Christmas post is probably a major challenge and it's pretty much a big thing mostly in English speaking countries. There are far less cards sent in say France for example.

Impetus

a65b2cd wrote: »

Interesting article - scanning technology has improved greatly since the late 1990s when that study was done.

The use of a reduced character set for the Eircode (around 25 instead of 36 letters and numbers) will make it very machine and human friendly.

Is the reference to a large volume of hand-written addresses from the 1990s as well - I suppose it is mainly Christmas cards?

I notice a huge difference in delivery performance between handwritten and well computer printed address. By well I mean monospaced characters (not proportional), putting the entire delivery address in upper case, etc and not using coloured envelopes. Much has to do with the quality of the handwriting, with is not improving with the increasing use of computers. I get a lot of postal stuff arising from me making a phone call, and this tends to be hand written addresses. Aside from call centres where they usually have a system to produce and package junk mail.

I know somebody who posted about 100 christmas cards to friends in Ireland in 2012, and most were delivered in January - despite her posting them in late November. And she probably has some of the clearest handwriting one will come across.

The two main sources of automation for postal and package processing (eg stuff used by DHL, Fedex etc) and Germany and Japan. There is no market for improving recognition accuracy of alpha numeric codes in the logistics processing machinery business. This is why handwritten letters take longer when posted to or in Ireland - there is no numeric postcode that can get the envelope moving in the system until it waits in a queue for some human to read and code it. The Germans (ie a division of Siemens) has 99.99% accuracy on recognition of all number postcoded addresses - which is also on the addresses used by perhaps 97% of the mail posted in the world. Only Britain, Malta and Canada use alpha. And Malta postal sorting is not mechanised. Japanese mail is sorted via the all numeric Japanese postcode. Because the Japanese also use 0123456789. It is the only element of an address in Japanese that a foreigner can generally read (in Kanji) - ie the postcode.

I suspect that many important issues have been ignored when deciding on the Eircode. The people involved don't care or are clueless. Which is what one gets with a government infrastructure that is not focused on providing a service to the customer. The "health service" is just one area that loudly manifests itself in the media - but the same lack of attention to quality of service is apparent through most activities that the government has a finger in the pie of.

SpaceTime

Christmas card issues won't really be resolved by postal codes. The issues it that you can't really design a system to cope with a huge spike in demand that lasts for a few weeks a year.

All postal services in any market where there's a tradition of sending xmas cards seem to suffer from it.

I guess the simplest way of ensuring your cards are delivered on time might be to do print the labels. I know one person who does that with her personal cards every year as she sends hundreds and it saves her a lot of hand writing i.e. just do a mail merge label production run on Word from a spreadsheet of addresses. Stick them on envelopes and then handwrite the cards.

Impetus

SpaceTime wrote: »

Christmas card issues won't really be resolved by postal codes. The issues it that you can't really design a system to cope with a huge spike in demand that lasts for a few weeks a year.

All postal services in any market where there's a tradition of sending xmas cards seem to suffer from it.

I guess the simplest way of ensuring your cards are delivered on time might be to do print the labels. I know one person who does that with her personal cards every year as she sends hundreds and it saves her a lot of hand writing i.e. just do a mail merge label production run on Word from a spreadsheet of addresses. Stick them on envelopes and then handwrite the cards.

While I agree mail-merge and labels is best, if a handwritten numeric postal code is machine readable, as it is in 99%+ of cases, the envelope gets scanned and partially OCR recognised - ie the 4 or 5 figure postcode. The envelope gets its usual barcode serial number and enters the system. The system can route the envelope, in the absence of being able to auto-recognise the entire address, using the postcode. This allows it to make its way to the destination delivery office before human intervention takes place. The human intervention can take place overnight (ie someone reading a scan of the address area of the envelope) on a VDU while the envelope is in a truck or in a box in a sorting office.

Once the human has defined the precise destination to the system, it can be auto sorted at the destination sorting office. Remember there is a large logistical issue in play - postal packets being collected in the afternoon from thousands of letterboxes etc and sent to one of four mechanized sorting offices. If they are non "local" they get routed to the next mechanized sorting office. And the same thing applies to mail to and from other countries using the same system.

Numeric postcodes would reduce the volume of envelopes etc requiring human intervention (thanks to the check-digit effect) of matching readable portions of the address and confirming this with the highly readable numeric postcode. The same considerations apply package delivery (eg of items ordered on the internet) - though in different ways.

SPDUB

Impetus wrote: »

I know somebody who posted about 100 christmas cards to friends in Ireland in 2012, and most were delivered in January - despite her posting them in late November. And she probably has some of the clearest handwriting one will come across.

If they were posted from Britain it sounds like she didn't put enough postage on them for delivery to ROI

Royal Mail regularly holds mail for insufficient postage to ROI , but with no return address in the UK , for about a month then sends it to An Post

Impetus

SPDUB wrote: »

If they were posted from Britain it sounds like she didn't put enough postage on them for delivery to ROI

Royal Mail regularly holds mail for insufficient postage to ROI , but with no return address in the UK , for about a month then sends it to An Post

They weren't mailed in GB. They were properly stamped and addressed.

You must work for gov.ie..... - treating the public like idiots mode seems to be one of their favourite occupations.

SPDUB

Impetus wrote: »

They weren't mailed in GB. They were properly stamped and addressed.

You must work for gov.ie..... - treating the public like idiots mode seems to be one of their favourite occupations.

You managed to ignore the word IF in your effort to take offence and throw around accusations

Jayuu

The amount of hyperbole on this thread is amusing. Apparantly seven months before it's introduced it's "the most hated post-code" on the planet. It apparantly is going to cause pensioners to die from heart attacks, delay everybody's Christmas cards, make it impossible for deliveries and also allow revenue to charge everybody millions of euro.

Personally I think the 99.9999999% of the people of the country don't really care about post codes and when they are introduced will slowly start to use them. Practically every computer system I know has an area for post codes so over a period of time recording and usage will increase, people will eventually become familiar with them and start to use them and you know the sky just might not fall down and the planet won't suddenly stop spinning.

I think a little reality check is called for on behalf of some people here.

Sam Russell

Well you are probably right.

However,

Computer voting - how is that going here, and did it get us results quicker?

PayPars, is that working well? Are all the health workers happy with it?

Rural Broadband? Is that roll out going well? It was Three that got the contract, how is that going?

HSE, how did that work out? Did it reduce costs and increase efficiency?

As for Irish Water ................

etc. etc. etc.

Impetus

The problem with the Eircode and all its complexity is that it does not provide a benefit to the customer – ie the public. It is not providing an address that one can use like 45 Aylesbury Road D4. The chances of it being widely used are minimal.

Eircode does not help you find a rural address any easier – you will invariably have to ask locals for directions – which is a primitive and time-wasting way of getting around efficiently, especially when making deliveries.

The creation of an Eircode will incur all the expense of giving each house a street/road address and building number – with no human readable output. They are effectively super-encrypting each address, with compression, to turn something like 45 Aylesbury Road Dublin 4 into something like D04 T8WX. To make things worse, 47 Aylesbury Road next door might look like D04 G9AB.

It is going to cost a fortune to develop and manage and keep up to date, and make sure that all users are up to date. And if you are building a new house or business premises – you will need the Eircode before the first brick is laid, to facilitate the delivery of building materials. How well will that work?

So on the benefits side it is almost zero benefit for the public and businesses.

On the cost side, it will probably be the most expensive “postcode” in the world. One of the newspapers today (SBP) reports that Irish Water’s operating expenses are twice the costs incurred by regional water companies in Britain. I predict that the Eircode costs will be a multiple of the costs paid by any other postal administration in Europe to create and maintain a postcode system for the country.

Which is typical of the performance of the Irish public service going back as far as I can remember. If there is a wrong way to do it, a committee of “public servants” will work the thinking through to maximise the “wrongness” for their own ends. To hell with the public.

This attitude is written all over the Eircode.

ozmo

Sam Russell wrote: »

Computer voting - how is that going here, and did it get us results quicker?

PayPars, is that working well? Are all the health workers happy with it?

Rural Broadband? Is that roll out going well? It was Three that got the contract, how is that going?

HSE, how did that work out? Did it reduce costs and increase efficiency?

As for Irish Water ................

etc. etc. etc.

There is a long list of cancelled ideas-
Some bad,( some good)- its good to know that bad ideas can be reversed no matter how far they have gone..

Government Decentralisation.. (Canceled)

Universal health Insurance (delayed maybe canceled)

West link toll bridge (bought back and updated)

Bertiebowl

Childrens hospital location

Still a chance if enough GIS using companies speak up it will get the professional review that was promised.

SpaceTime

What will happen is it'll be rolled out and not used.

I can't even imagine An Post will use it if it means a huge IT upgrade for no benefit.

TheBustedFlush

Impetus wrote: »

The problem with the Eircode and all its complexity is that it does not provide a benefit to the customer – ie the public. It is not providing an address that one can use like 45 Aylesbury Road D4. The chances of it being widely used are minimal.

Eircode does not help you find a rural address any easier – you will invariably have to ask locals for directions – which is a primitive and time-wasting way of getting around efficiently, especially when making deliveries.

The creation of an Eircode will incur all the expense of giving each house a street/road address and building number – with no human readable output. They are effectively super-encrypting each address, with compression, to turn something like 45 Aylesbury Road Dublin 4 into something like D04 T8WX. To make things worse, 47 Aylesbury Road next door might look like D04 G9AB.

It is going to cost a fortune to develop and manage and keep up to date, and make sure that all users are up to date. And if you are building a new house or business premises – you will need the Eircode before the first brick is laid, to facilitate the delivery of building materials. How well will that work?

So on the benefits side it is almost zero benefit for the public and businesses.

On the cost side, it will probably be the most expensive “postcode” in the world. One of the newspapers today (SBP) reports that Irish Water’s operating expenses are twice the costs incurred by regional water companies in Britain. I predict that the Eircode costs will be a multiple of the costs paid by any other postal administration in Europe to create and maintain a postcode system for the country.

Which is typical of the performance of the Irish public service going back as far as I can remember. If there is a wrong way to do it, a committee of “public servants” will work the thinking through to maximise the “wrongness” for their own ends. To hell with the public.

This attitude is written all over the Eircode.

Don't agree with this thinking - it's flawed.

ozmo

SpaceTime wrote: »

What will happen is it'll be rolled out and not used.

I can't even imagine An Post will use it if it means a huge IT upgrade for no benefit.

I would say AnPost will be forced use it- even if it cost millions to update the ocr equipment they have already. Same with Garda and anyone else currently paying for geodirectory.

If there is a licence payment of even 50 to 100 per year(never mind the "something more than 5000" the website currently quotes) to get an addon for tomtom to convert postcode to a location- I doubt any but delivery companies and gov offices will buy it.

I would imagine very many will add it as part of their address without thinking- but without paying for a licence it will be of less use to the general public than it could have been. Wasted opportunity.

TheBustedFlush

ozmo wrote: »

I would say AnPost will be forced use it- even if it cost millions to update the ocr equipment they have already. Same with Garda and anyone else currently paying for geodirectory.

If there is a licence payment of even 50 to 100 per year(never mind the "something more than 5000" the website currently quotes) to get an addon for tomtom to convert postcode to a location- I doubt any but delivery companies and gov offices will buy it.

I would imagine very many will add it as part of their address without thinking- but without paying for a licence it will be of less use to the general public than it could have been. Wasted opportunity.

An Post have already stated that they will be using it and what it will cost them to update software/systems (about a million??). There's an article somewhere on the web quoting them.

TomTom already pay licence fees to use other postcode/address systems around the world that contain separate data around postcodes and geocoordinates - Ireland will be no different in that regard. It's a question of how much they can negotiate the cost for. Google will do the same. As will Nokia's company Here - which supplies Garmin, some in-car systems and etailers like Amazon.

The public won't need a licence since there will be a free public website on which you can lookup and get postcodes/addresses and see where they are.

State bodies will obviously be using them as well since they're being put onto their databases. So will banks, insurance companies, e-tailers, etc who see the likely cost-savings/fraud prevention as a benefit to them, and who may also see a better customer service provision also. Some may do, some may not.

Where it may fall down initially is if people don't voluntarily provide it when asked. Eventually, we may get to the same place that other countries have where you're asked for it automatically as part of engaging with businesses for queries or supply of services.

Sam Russell

ozmo wrote: »

There is a long list of cancelled ideas-
Some bad,( some good)- its good to know that bad ideas can be reversed no matter how far they have gone..

Government Decentralisation.. (Canceled)

Universal health Insurance (delayed maybe canceled)

West link toll bridge (bought back and updated)

Bertiebowl

Childrens hospital location

Still a chance if enough GIS using companies speak up it will get the professional review that was promised.

The list I gave was of unmitigated disasters that cost millions.

The West Link cost £10 million to build and was sold back to the state for €500 million. The operators refused to open the gates to prove that they were the cause of all the delays on the M50. There are no delays on the M50 now.

The Children's Hospital is due to open when?

Government de-centalisation was a daft idea. It sort of happened, but not in a good way.

Now to get back to postcodes:-

It would be possible to reverse the postcode project, but it would be through a redesign along the lines debated here.

1. Numeric codes

2. The codes to be 4 or 5 digit code to begin with and possible hidden extra numbers to give location down to building

3. All properties to have actual addresses including house/building numbers.

The charges for the system to be zero for most users.

TheBustedFlush

Sam Russell wrote: »

The list I gave was of unmitigated disasters that cost millions.

The West Link cost £10 million to build and was sold back to the state for €500 million. The operators refused to open the gates to prove that they were the cause of all the delays on the M50. There are no delays on the M50 now.

The Children's Hospital is due to open when?

Government de-centalisation was a daft idea. It sort of happened, but not in a good way.

Now to get back to postcodes:-

It would be possible to reverse the postcode project, but it would be through a redesign along the lines debated here.

1. Numeric codes

2. The codes to be 4 or 5 digit code to begin with and possible hidden extra numbers to give location down to building

3. All properties to have actual addresses including house/building numbers.

The charges for the system to be zero for most users.

The last line is the important one. There's no such thing as a free lunch. Who's going to pay for this - the taxpayer is who. When you say "most users", who do you think should pay?

Impetus

TheBustedFlush wrote: »

Don't agree with this thinking - it's flawed.

Well perhaps sometime you might spell out the "flaws".

Impetus

TheBustedFlush wrote: »

The last line is the important one. There's no such thing as a free lunch. Who's going to pay for this - the taxpayer is who. When you say "most users", who do you think should pay?

The citizen will pay. And will get zero in return. A wasted opportunity to bring Ireland into the 21st century. Presided over by successive government ministers over the past five years + and the dysfunctional permanent government.

I don't believe the last line is the most important one (costs). The previous line is. ie "3. All properties to have actual addresses including house/building numbers." To which I would add a standardised town name. Give that town a number, and if it is big, multiple numbers pointing to the various zones centred on it. Every property already has a unique code in the Geodirectory. The problem with the latter is that it is unwieldy and needlessly complex in its structure, and does not assign road numbers and building numbers in rural areas. That is what Capita should be working on. Give people value for the money they are spending.

TheBustedFlush

Impetus wrote: »

The citizen will pay. And will get zero in return. A wasted opportunity to bring Ireland into the 21st century. Presided over by successive government ministers over the past five years + and the dysfunctional permanent government.

You've picked this up wrongly. I was responding to Sam Russell's proposed alternative solution based on numeric codes and his statement that: "The charges for the system to be zero for most users."

Impetus

TheBustedFlush wrote: »

You've picked this up wrongly. I was responding to Sam Russell's proposed alternative solution based on numeric codes and his statement that: "The charges for the system to be zero for most users."

The cost won't be zero for most users. They may not get an invoice for a "postcode levy" to use the dumb terminology one often hears. The cost will comprise of things like needlessly inflated delivery costs for stuff ordered on the internet etc as truck drivers try to find the house in question. And similar in many other activities involving search of a premises. Anything involving logistics to business or domestic premises.

This is a missed opportunity for an infrastructure improvement that could add 3 to 5% to the Irish GDP. Which is a huge cost to society, which will be replicated over and over ad infinitum.

One can achieve twice as much in a country with a good infrastructure. ie It is far easier for people to be more productive and increase their and the country's national wealth. Ireland has been going in the other direction for decades.

TheBustedFlush

Impetus wrote: »

The cost won't be zero for most users. They may not get an invoice for a "postcode levy" to use the dumb terminology one often hears. The cost will comprise of things like needlessly inflated delivery costs for stuff ordered on the internet etc as truck drivers try to find the house in question. And similar in many other activities involving search of a premises. Anything involving logistics to business or domestic premises.

This is a missed opportunity for an infrastructure improvement that could add 3 to 5% to the Irish GDP. Which is a huge cost to society, which will be replicated over and over ad infinitum.

One can achieve twice as much in a country with a good infrastructure. ie It is far easier for people to be more productive and increase their and the country's national wealth. Ireland has been going in the other direction for decades.

And you're still on the wrong track. I was asking Sam Russell to answer the question about who would pay for his proposed solution. Not sure why you're getting involved on that at all, since you didn't propose the solution.

Sam Russell

TheBustedFlush wrote: »

You've picked this up wrongly. I was responding to Sam Russell's proposed alternative solution based on numeric codes and his statement that: "The charges for the system to be zero for most users."

The cost would be zero for most users for a properly designed post code system. If a 4 or 5 dgit code was used, it would cost zero once the code was designed. Most users would find that sufficient and so not want the premium (hidden) last 4 or 5 digits to identify down to the building. It is these last digits that cost the money to maintain.

Using numbers reminiscent of the phone numbers (1 for Dublin 21 for Cork, 91 for Galway, 51 for Waterford, etc.) that is already there, and since the numbers are allocated by telephone exchanges, they are roughly on population density, the 5 digit code would give about 300 or so houses/buildings per code. (1.5 million houses divide by 5,000).

The house number would reduce this to maybe one house, and together with the street name to definitely one house. New house gets the same post code as neighbour.

No cost once established except for the hidden code and that should be paid for by the users of the hidden part.

.