National Postcodes to be introduced

Sam Russell

There is a very good case for having post codes - it is just that this particular design misses most of the requirements.

The way the current proposal/design is constructed is poorly executed. The first 3 characters defining a postal town is a basic folly because these could change into the future, and they leave resulting routing areas as huge with up to 80,000 addresses in the largest of them. The routing code should have defined a much smaller area of about 200 addresses which could have been achieved by a five digit numeric code, or a four character alpha-numeric code. Numbers have the benefit of not producing rude codes, and are harder to make them into ghettos.

Another objection is that it will turn into a revenue stream for a private company with no proper state supervision. So it not only costs the state €25m, it costs Irish business and consumers a similar sum. It should be fully at the states expense.

How can other countries provide a better system for free?

SpaceTime

Area code snobbery in London in the early 1990s
You can see why they were terrified of touching the Dublin district numbering.

https://www.youtube.com/watch?v=1RQh9ijeVJ0

antoinolachtnai

Not really. The problem in Ireland, as has been stated previously, is that stating the name of your locality is basically meaningless. 'Warrenstown, Co Meath' just doesn't refer to anything in particular. It may refer to up to 5 different places. 'Malahide, Co. Dublin' is similarly a text without a well defined referent. Even at the more local level, for things like 'Sandymount, Dublin 4', there is a great deal of ambiguity and interpretation. There are thousands of similar cases. It is a term that can be construed many different ways and the total area which it refers to is enormous.

Is it really too much to expect the postcode to resolve this problem? Any reasonably standard postcode model from around the world would have dealt with this to a greater or lesser extent.

If it is really so important to avoid postcode ghettoes, then why not randomize the whole code? Not just the last four characters?

Leonard Shelby

antoinolachtnai wrote: »

Is it really too much to expect the postcode to resolve this problem? Any reasonably standard postcode model from around the world would have dealt with this to a greater or lesser extent.

None of which would work for Ireland. To solve non-unique addressing requires a unique postcode. Looking for hierarchy in the postcode to solve privacy concerns completely misses the point. It isn't the infrequent times you don't want to provide your full postcode to a website that's the problem (give them the postcode of your local post office instead) it is the potential misuse of full postcodes stored in address databases. That concern is dealt with by enforcing data protection legislation.

If it is really so important to avoid postcode ghettoes, then why not randomize the whole code? Not just the last four characters?

Because then it wouldn't be a postcode. Routing Keys are required to manually sort post that hasn't been read by OCR.

antoinolachtnai

Leonard Shelby wrote: »

None of which would work for Ireland. To solve non-unique addressing requires a unique postcode.

There is no problem having a unique postcode that is hierarchical. There is no opposition whatsoever between these two objectives.

The current solution doesn't solve the non-unique addressing problem. It just spreads random number pixie dust across the country, ignores the fundamental problem and hopes it will go away.

Looking for hierarchy in the postcode to solve privacy concerns completely misses the point.

Why is designing privacy into the code missing the point? A system which is so orthogonal to international best practice is introducing massive uncertainty in all sorts of ways.

It isn't the infrequent times you don't want to provide your full postcode to a website that's the problem (give them the postcode of your local post office instead) it is the potential misuse of full postcodes stored in address databases. That concern is dealt with by enforcing data protection legislation.

So privacy is an afterthought? Who is going to do all this enforcing? Who is going to pay for it? Is there going to be a levy on the postcode?

I asked why the whole postcode shouldn't just be randomised, not just the second half, if it is so important to avoid postcode ghettoes.

You answered:

Because then it wouldn't be a postcode. Routing Keys are required to manually sort post that hasn't been read by OCR.

If this is true, your advocacy for randomising the second part of the code doesn't make a lot of sense. If routing keys are so absolutely essential and useful for sorting at post town level, even in this modern computerised era, then why not continue them on to allow manual sorting at local delivery level? How would this be at all harmful?

ukoda

antoinolachtnai wrote: »

How would this be at all harmful?

Because now you've to group houses and this where people willbe up in arms being grouped with "undesirable" houses

SpaceTime

ukoda wrote: »

Because now you've to group houses and this where people willbe up in arms being grouped with "undesirable" houses

People need to get over themselves and politicians should stop pandering to this kind of extreme snobbery.

Even Dublin 4 isn't all wealthy and includes some quite historically deprived areas.

Leonard Shelby

antoinolachtnai wrote: »

The current solution doesn't solve the non-unique addressing problem. It just spreads random number pixie dust across the country, ignores the fundamental problem and hopes it will go away.

It appears you don't understand the postcode design.

Why is designing privacy into the code missing the point? A system which is so orthogonal to international best practice is introducing massive uncertainty in all sorts of ways.

Best practice is to implement a design that works, not copy one that won't. You appear to have difficulty understanding this concept.

So privacy is an afterthought? Who is going to do all this enforcing? Who is going to pay for it? Is there going to be a levy on the postcode?

Privacy isn't ensured by having a hierarchical postcode, that is complete nonsense. Are you going to ensure that people only provide partial postcodes at all possible times? Data protection laws already exist and are enforced, you do know this don't you?

If this is true, your advocacy for randomising the second part of the code doesn't make a lot of sense. If routing keys are so absolutely essential and useful for sorting at post town level, even in this modern computerised era, then why not continue them on to allow manual sorting at local delivery level? How would this be at all harmful?

Because they aren't required, would add additional cost, require to be changed with new developements, cause unfair social consequences, etc. etc. I appreciate that you lack sufficient knowledge in this area to understand these issues, but that doesn't translate into legitimate concerns. You would be better advised researching the area rather than displaying your ignorance. Until you demonstrate minimal competence in this area I'm not going to bother to continually repeat myself.

ardmacha

Postcodes do not cause unfair social consequences, they only allowed identification of areas.Randomization prevents this, but also prevents a lot of positive uses.

Carawaystick

Leonard Shelby wrote: »

Data protection laws already exist and are enforced, you do know this don't you?

Ah Here.
british intelligence are copying all irish internet traffic,
the cops were recording calls to cop stations,
irish water were demanding ppsn without any basis.
Alan Shatter and Michael McDowell used info collected as part of DoJ umbrella for electioneering


No-one will be prosecuted for these data protection breaches. Shatter may get a civil slap in the wrist.

antoinolachtnai

Leonard Shelby wrote: »

It appears you don't understand the postcode design.

I understand it fine. Unlike some people, I have taken the trouble to read the project documentation. Aside from the problems with the design itself, I just can't understand how it is going to be brought into operation for less than five million euros above the costs that have already been made public.

Best practice is to implement a design that works, not copy one that won't. You appear to have difficulty understanding this concept.

This is not my language. The project documentation points out that the design is not international best practice.

You would be better advised researching the area rather than displaying your ignorance. Until you demonstrate minimal competence in this area I'm not going to bother to continually repeat myself.

It would be better if you explained your position rather than retreating to personal insults against me and dismissing my concerns out of hand. You are building 'straw man' versions of my points in order to dismiss them, rather than considering them on their merits.

I can see from your posting history since October 2013 that you specifically advocate this scheme and it would appear that that is what you are here to do. That is fair enough, I suppose, but now I find you are willing to broach no criticism of the eircode scheme.

You seem to think I have no expertise in this area; in fact I am a director of a national NGO in the privacy area and have fought and won on privacy issues in the European Court. Because of my expertise (I suppose), the Department of Communications, Energy and Natural Resources recently met with me to consider concerns our organization had raised. I am surprised that you are simply dismissing my concerns out of hand, because that has not by any means been the attitude of the Department. Separately, I am also the convener of a national level committee on character sets and encoding.

Anyone reading this thread should remember that there is a lot at stake for the people who are promoting this code, and there are a lot of people who have to protect themselves. I have looked at this every way, and from what I can see, the design of this scheme has less to do with post coding concerns and more to do with protecting intellectual property. The design of this code is specifically to make sure that the postcode cannot be 'opened' in the way the UK postcode was opened by campaigners.

The merits of an 'open' postcode are not considered at all anywhere in the design documentation for the code and this should be a major concern for all concerned. The promoters of this code want to shut down debate and silence critics. This is not a good thing.

Leonard Shelby

antoinolachtnai wrote: »

This is not my language. The project documentation points out that the design is not international best practice.

I assume you are referring to the pros and cons page that has been tweeted recently. The phrase is "not currently international best practice", and the word currently is important. This isn't a red mark against the design, it is simply stating that this is a new design and therefore can't adhere to current best international practice.

It would be better if you explained your position rather than retreating to personal insults against me and dismissing my concerns out of hand. You are building 'straw man' versions of my points in order to dismiss them, rather than considering them on their merits.

I can see from your posting history since October 2013 that you specifically advocate this scheme and it would appear that that is what you are here to do. That is fair enough, I suppose, but now I find you are willing to broach no criticism of the eircode scheme.

You seem to think I have no expertise in this area; in fact I am a director of a national NGO in the privacy area and have fought and won on privacy issues in the European Court. Because of my expertise (I suppose), the Department of Communications, Energy and Natural Resources recently met with me to consider concerns our organization had raised. I am surprised that you are simply dismissing my concerns out of hand, because that has not by any means been the attitude of the Department. Separately, I am also the convener of a national level committee on character sets and encoding.

Anyone reading this thread should remember that there is a lot at stake for the people who are promoting this code, and there are a lot of people who have to protect themselves. I have looked at this every way, and from what I can see, the design of this scheme has less to do with post coding concerns and more to do with protecting intellectual property. The design of this code is specifically to make sure that the postcode cannot be 'opened' in the way the UK postcode was opened by campaigners.

The merits of an 'open' postcode are not considered at all anywhere in the design documentation for the code and this should be a major concern for all concerned. The promoters of this code want to shut down debate and silence critics. This is not a good thing.

The decision whether the postcode is open or not is purely a decision for Government. If they decided to pay for this from general taxation rather than a business levy then the design won't stop them from doing so.

I didn't make any personal insults. I assessed your competence based on your posts, if you felt insulted by my assessment that is your problem. Please don't trot out an invalid "argument from authority" you have to demonstrate competence by the strength and validity of your arguments.

plodder

ukoda wrote: »

You can still give your locality. No one had removed that option

In the UK postcode, you have for instance:

BT - is the whole of Northern Ireland

BT74 - is Enniskillen and its surrounding area.

BT74 8 - is some area within the Enniskillen area

BT74 8XY is then some street level area within the above.

So, what you're suggesting with Eircode (examples made up)

W09 - refers to Naas (say)

W09-XP19 is a random address in the Naas postal area

So, if I want to refer to a specific area near Naas (eg Waterstown), then what do I use? The XP19 is no use. So, you have to drop that.

W09/Waterstown for example? How do I know there is only one Waterstown in the Naas postal district? We already know that townland names are not unique in the same county.

So, your suggestion does not solve the problem, or provide any improvement on what we have currently without postcodes.

Richard

plodder wrote: »

In the UK postcode, you have for instance:

BT - is the whole of Northern Ireland

BT74 - is Enniskillen and its surrounding area.

BT74 8 - is some area within the Enniskillen area

BT74 8XY is then some street level area within the above.

And coupled with a house number (or name) you have the same level of detail as Eircode with the flexibility of an open system.

antoinolachtnai

Leonard Shelby wrote: »

The decision whether the postcode is open or not is purely a decision for Government. If they decided to pay for this from general taxation rather than a business levy then the design won't stop them from doing so.

A business levy?

You know very well that you trotted out personal insults. You accused me of 'displaying my ignorance', for instance. That is an insult. You decided it was easier to trot out these insults instead of dealing with the issue, viz. that eircode has been designed to protect intellectual property rather than to be useful, that there are major privacy disadvantages to eircode compared to best practice postcode systems and that eircode is going to be extremely expensive, if not impossible to implement because of the flawed design.

Seeing as you consider yourself an expert on the topic, it would be welcome if you would now put forth your credentials. But I do not expect that you will. It would also be welcome if you apologised for your insulting remarks about me, but again, I do not expect you will.

OssianSmyth

I've uploaded a copy of the Eircode Database technical specification for anyone interested.

Leonard Shelby

antoinolachtnai wrote: »

A business levy?

You know very well that you trotted out personal insults. You accused me of 'displaying my ignorance', for instance. That is an insult. You decided it was easier to trot out these insults instead of dealing with the issue, viz. that eircode has been designed to protect intellectual property rather than to be useful, that there are major privacy disadvantages to eircode compared to best practice postcode systems and that eircode is going to be extremely expensive, if not impossible to implement because of the flawed design.

Ignorance is defined as a lack of knowledge or information.

Eircode has not been designed to protect intellectual property rather than to be useful. That is an ignorant comment. Claiming that it is going to be extremely expensive if not impossible to implement falls into the same category.

You are making baseless claims. You really can't see any other reasons why hierarchy isn't built into the postcode other than to protect IP? You don't understand how An Post will deliver the postcode to every house so you pluck a figure of 5 million from the air and present it as a reasoned argument? And you wonder why I don't take your arguments seriously?

Leonard Shelby

antoinolachtnai wrote: »

Seeing as you consider yourself an expert on the topic, it would be welcome if you would now put forth your credentials.

Credentials do not make invalid arguments valid. You've proven that.

antoinolachtnai

You haven't explained what the 'business levy' will be. Is it a new type of tax?

You should make us more ignorant, by disclosing all the information you have instead of selectively disclosing it when it suits you. You will not, of course, because this whole project is characterised by cover-up and secrecy. For instance, there is an 804 page contract for delivery of the eircode, but it turns out that there is not a single word of this which is not confidential.

The whole eircode idea hinges on An Post being able to deliver a 'postcode notification item' 99.9999 percent accurately through the regular post. The reality is that this is impossible to do through the postal network. I have given detailed figures showing why this is impossible, and you have made no response other than to say that An Post must have some way or other to do it. But they haven't, other than to do something very specific and bespoke for eircode

My view, yes, is that it would cost at least 5 million euros to roll out that specific and bespoke delivery program for eircode because of the inherent flaw in the design and that this cost in addition to the budgeted amounts disclosed to the Oireachtas committee. It should be pretty obvious where that cost comes from but if you can't work it out, I can go back and spell it out for you.

These are all claims based on facts and reasonable estimates. You dispute these facts and estimates but you aren't putting forward any alternatives. You just claim that everything is going to be fine and you expect the public to accept that. It won't. The eircode program is just planning to ignore the problem and hope for the best. The result will be a shoddy system which will take years to fix.

yuloni

This post has been deleted.

alistair spuds

The state are waiting to figure out a way to award it to one of Denis O'Brien's concerns

Leonard Shelby

antoinolachtnai wrote: »

You haven't explained what the 'business levy' will be. Is it a new type of tax?

Eircode is free for personal use. People will look it up on on their website, or in Google maps etc. for free. The only charge is for commercial use, businesses that will benefit financially due to decreased costs or increased revenue will be charged a fee. If neither apply then businesses won't licence Eircode if it provides no positive cost benefit.

The whole eircode idea hinges on An Post being able to deliver a 'postcode notification item' 99.9999 percent accurately through the regular post. The reality is that this is impossible to do through the postal network. I have given detailed figures showing why this is impossible, and you have made no response other than to say that An Post must have some way or other to do it. But they haven't, other than to do something very specific and bespoke for eircode

99.9999 must be a figure you've made up, as it isn't one that would be the outcome of an exercise in determining acceptable delivery error rate. If 2.2 million postcodes are being delivered then the exercise would be a failure by your definition if 3 postcodes were delivered to the wrong address. 99.95% would yield 1,100 errors, which would be an acceptable error rate that can be easily corrected. If only 99.5% accuracy was achieved then the correction of 11,000 errors would indeed be an onerous exercise.

My view, yes, is that it would cost at least 5 million euros to roll out that specific and bespoke delivery program for eircode because of the inherent flaw in the design and that this cost in addition to the budgeted amounts disclosed to the Oireachtas committee. It should be pretty obvious where that cost comes from but if you can't work it out, I can go back and spell it out for you.

So if An Post have a plan that will ensure delivery to within an acceptable error rate within the current budget your concerns disappear?

These are all claims based on facts and reasonable estimates. You dispute these facts and estimates but you aren't putting forward any alternatives. You just claim that everything is going to be fine and you expect the public to accept that. It won't. The eircode program is just planning to ignore the problem and hope for the best. The result will be a shoddy system which will take years to fix.

Using terms like "shoddy system" doesn't enhance your argument. The public are being told that they will receive their postcode next year, and that using it will be in their best interests. None of the benefits for the public require a hierarchy in the postcode. The design assumes availability of modern technology, same as when you want to find somewhere using Google. Unless the postcode is going to be accompanied by a massive re-signage project for the country (completely impractical) or a re-addressing project (wait 20 years for that to play out) then a hierarchy is of absolutely no value.

antoinolachtnai

Leonard Shelby wrote: »

Using terms like "shoddy system" doesn't enhance your argument. The public are being told that they will receive their postcode next year, and that using it will be in their best interests.

What has that got to do with anything? Eircode is a shoddy design for a postcode system. We have been through these points before, but let me tell you the ways:

1. The whole design is contrary to international best practice.

2. There has been no independent oversight or evaluation of the eircode project as a whole. (This is a fact accepted by the Department.)

3. There is no implementation plan to speak of.

4. It is closed where it could be open. In fact it has been specifically designed so that it will be impossible to open up without violating the licence holder's IPR.

5. The whole thing is a big commercial mystery. Every word of every page of the 804-page contract between DCENR and Capita for the eircode is being treated as a state secret.

6. The revenue model is broken. You have provided indicative pricing here, the first time such pricing information has been made available to the public. Based upon that pricing, eircode will never be able to recover the investment.

7. The pricing model will make it unviable for civic projects like kildarestreet.com

8. There is massive uncertainty about the privacy problems that the unique design of the code will bring.

The fact that the Product Guide for the product has been withdrawn for the moment and that the eircode sales team are no longer responding to queries about it is indicative of the situation.

Eircode is free for personal use. People will look it up on on their website, or in Google maps etc. for free. The only charge is for commercial use, businesses that will benefit financially due to decreased costs or increased revenue will be charged a fee. If neither apply then businesses won't licence Eircode if it provides no positive cost benefit.

Non-profits that want to use eircode for civic purposes (the obvious one is kildarestreet.com but there are many more) will have to pay full whack. This is not the case in the UK or elsewhere.

99.9999 must be a figure you've made up, as it isn't one that would be the outcome of an exercise in determining acceptable delivery error rate. If 2.2 million postcodes are being delivered then the exercise would be a failure by your definition if 3 postcodes were delivered to the wrong address. 99.95% would yield 1,100 errors, which would be an acceptable error rate that can be easily corrected. If only 99.5% accuracy was achieved then the correction of 11,000 errors would indeed be an onerous exercise.

According to the documentation, one of the key purposes of eircode is as a means to direct emergency medical services to the correct location. To make this work, all the codes have to be right, not just the ones that are frequently used and can be corrected.

6-nines is a pretty standard figure for accuracy in high quality manufacturing and in some parts of the services sector. The reason I put it forward here is that the accuracy of the initial delivery of postcode notification items is so critical. You think that it would be straightforward to track down 1,100 errors. I canot see how this could be done easily. A wrong eircode could lie undetected for years and could cause a disaster when it reared its head in an emergency situation.

If the code was hierarchical, this would not be so critical. A misdirected ambulance would probably be at least close to the desired destination. It would also be much easier for postmen and others to spot postcode notification items which had gone to completely the wrong address. With the crazy random scheme it is absolutely critical that every code is delivered to the correct house the first time. It will be very difficult to detect problems retrospectively.

So if An Post have a plan that will ensure delivery to within an acceptable error rate within the current budget your concerns disappear?

It really depends on what the plan is. There are lots of other problems with this code, including the 'over-specificity' privacy problem (though the 'bootstrapping' problem also raises major issues of privacy for people who are given the wrong eircodes and have their mail mixed up with someone else's.)

None of the benefits for the public require a hierarchy in the postcode. The design assumes availability of modern technology, same as when you want to find somewhere using Google.

It might not require a hierarchy in theory, but in practice it would make it a lot easier to implement and use.

Eircode as it stands is just a big theoretical idea. No real case has been made that it can actually be implemented in practice with a reasonable level of accuracy.

Unless the postcode is going to be accompanied by a massive re-signage project for the country (completely impractical) or a re-addressing project (wait 20 years for that to play out) then a hierarchy is of absolutely no value.

That is a bit of a straw man argument.

Why would using the fourth and fifth characters of the code to refer to the Small Area require a massive re-signage project? Please be specific here (although I am pretty sure you will not be specific and will just ignore the possibility.)

Carawaystick

OssianSmyth wrote: »

I've uploaded a copy of the Eircode Database technical specification for anyone interested.

I notice page 3, section1.1 of
http://www.ossiansmyth.ie/wp-content/uploads/2014/10/ECAD-Product-Guide-Edition-1-V-1-7-Published-.pdf
states
Every address in Ireland will receive its Eircode by mid 2015.
This is much later than the previously stated end of April date.
(http://www.eircode.ie/residential/overview)

The other point is the bootstrapping issue.
How will capita get an Post to deliver letters to individual addresses, with out an Post using an post's knowledge of the adressees?
But more importantly, given that eircode.ie states clearly in the link above

Eircode.ie wrote:

You will be sent a letter

How will Capita or eircode.ie know who lives at an address?
If they don't know who lives there, how will they send everyone living there a letter?
If they do know everyone living there, how did they find that out?
If they meant to not send everyone living at every address a letter informing them personally of their eircode, they are lying.

SpaceTime

I'm guessing they'll send letters to "The Occupier", 1 Main Street, Anytown, Co. Anywhere.

However, in rural areas with non unique addresses that's not going to work as there mightn't be a house name. The only way of addressing the house might be by family name...

Leonard Shelby

antoinolachtnai wrote: »

What has that got to do with anything? Eircode is a shoddy design for a postcode system. We have been through these points before, but let me tell you the ways:

My oh my. You are hopelessly biased. Because the design impacts on privacy concerns as it is unique to each property you're simply incapable of assessing competing arguments correctly and assume that anything you don't understand or don't have enough information about is evidence in your favour. I recommend reading Daniel Kahneman's Thinking Fast and Slow as a way of enlightening yourself. John Tuohy from Nightline gave a very good presentation to the joint Oireachtas committee recently, that would be a good place to start to test your bias.


Let me respond to your "analysis"

1. The whole design is contrary to international best practice.

International best practice evolves. What was best practice twenty years ago has to be re-evaluated to determine if it still applies. You would like a hierarchy in the code for certain privacy concerns, completely ignoring the negative impact of such a design change which far outweigh the minimal benefits to which you allude.

2. There has been no independent oversight or evaluation of the eircode project as a whole. (This is a fact accepted by the Department.)

PA Consulting are performing this role.

3. There is no implementation plan to speak of.

Absence of evidence isn't evidence of absence. You haven't seen a plan, that doesn't mean it doesn't exist or isn't detailed. Nonsense.

4. It is closed where it could be open. In fact it has been specifically designed so that it will be impossible to open up without violating the licence holder's IPR.

All IPR resides with the State. The licence holder is a license holder to operate the postcode. Nonsense.

5. The whole thing is a big commercial mystery. Every word of every page of the 804-page contract between DCENR and Capita for the eircode is being treated as a state secret.

Shock horror, commercially sensitive contract signed with a PLC is deemed to be commercially sensitive. How naive are you?

6. The revenue model is broken. You have provided indicative pricing here, the first time such pricing information has been made available to the public. Based upon that pricing, eircode will never be able to recover the investment.

Is it prohibitively expensive or won't they charge enough to make it viable, make up your mind.

7. The pricing model will make it unviable for civic projects like kildarestreet.com

Wrong. You're relying on Loc8 propaganda for your pricing information.

8. There is massive uncertainty about the privacy problems that the unique design of the code will bring.

Which you have been completely unable to articulate. Because something that is illegal is easier to do today than it was yesterday doesn't make it less illegal.

According to the documentation, one of the key purposes of eircode is as a means to direct emergency medical services to the correct location. To make this work, all the codes have to be right, not just the ones that are frequently used and can be corrected.

6-nines is a pretty standard figure for accuracy in high quality manufacturing and in some parts of the services sector. The reason I put it forward here is that the accuracy of the initial delivery of postcode notification items is so critical. You think that it would be straightforward to track down 1,100 errors. I canot see how this could be done easily. A wrong eircode could lie undetected for years and could cause a disaster when it reared its head in an emergency situation.

If the code was hierarchical, this would not be so critical. A misdirected ambulance would probably be at least close to the desired destination. It would also be much easier for postmen and others to spot postcode notification items which had gone to completely the wrong address. With the crazy random scheme it is absolutely critical that every code is delivered to the correct house the first time. It will be very difficult to detect problems retrospectively.

Ok, I'll try to be kind as you obviously haven't thought this one through. The postcodes will be delivered by An Post using their postal delivery routes. The postcodes will be sorted by route delivery order. If a postcode is delivered to the wrong address it will be close to the desired destination. Sigh.

It might not require a hierarchy in theory, but in practice it would make it a lot easier to implement and use.

Simply wrong on both counts.

Eircode as it stands is just a big theoretical idea. No real case has been made that it can actually be implemented in practice with a reasonable level of accuracy.

There you go again...

Why would using the fourth and fifth characters of the code to refer to the Small Area require a massive re-signage project? Please be specific here (although I am pretty sure you will not be specific and will just ignore the possibility.)

The argument peddled for hierarchy and sequencing is it assists you in letting you know you are near the correct location. For this information to be useful without access to modern technology would require signage.

Small Areas are built into the database and are available in a simple lookup. Go talk to CSO and they can explain why they are not required to be in the actual postcode.

Please read the book before responding.

plodder

Leaving aside the unique addressing issue (which I think Eircode solves pragmatically) let's compare hierarchical codes against random codes in general.

Advantages of a random code

1. Avoids (to an extent) the postcode snobbery problem, but only to the extent that existing Dublin postcodes are established. The unfamiliar provincial posttown boundaries will still have this problem, especially at county boundaries. Also, people shouldn't be under any illusion that insurance companies still (and always have) rated risk based on areas and addresses. Hiding the issue doesn't make it disappear.

2. Hides the postcode "IP" so that it has to be licensed. However, it is arguable that even with an open hierarchical code, most large businesses would license it anyway, for address verification purposes.

Advantages of a hierarchical code

1. Easier to learn/remember. Structured information is always easier to remember than random information. Because hierarchical codes are more "recognisable" there is less scope for error when calling them out.

2. Small businesses are not likely to license the Eircode datasets. So, Eircode is of limited use to them. The typical example is a tradesman who has a list of callouts in a day and wants to organise his traveling time efficiently. With a hierarchical code, he can do that just by looking at the code. With Eircode, he will have to buy some third party product (assuming such a thing will exist) to disentangle the random codes. Why impose technology on people for its own sake (remember evoting!)

3. Data protection concerns. For some purposes (eg anonymous market surveys) a hierarchical code allows people to disclose where they live, but without disclosing an exact address. Not possible with Eircode.

4. New house builds. The one time when a householder most needs an accurate post(location) code, is when a new house gets built. It's likely to take months before an Eircode is allocated, reflected in a quarterly update, and then updated in third party applications. A hierarchical code will have some level of position accuracy from before a house gets built since the area (in most cases) pre-exists the house.

5. Better for emergency services. If a householder wants to call the emergency services to their area, because there is an incident somewhere nearby, but not at their own property they can't simply specify the area. They have to provide their own code, which could result in callout charges being directed to them.

6. Better for government. At least two government departments said they wanted a hierarchical code. I don't think any said they wanted a random code. The advantages are obvious in terms of a consistent view of where people live. Also, small areas can be aggregated in different ways. So, the postcode doesn't impose one particular structure for all purposes. Eg you can combine small areas in one way to make up electoral divisions, and in a different way to make up school districts.

Sam Russell

Any post code could address the unique address problem, just make the small area small enough.

It is still not too late to make Eircode hierarchical.

It can be restructured after launch if it is a disaster - it is only software.

SpaceTime

Sam Russell wrote: »

Any post code could address the unique address problem, just make the small area small enough.

It is still not too late to make Eircode hierarchical.

It can be restructured after launch if it is a disaster - it is only software.

I'd say it'll launch in 2032 at this rate...

Leonard Shelby

plodder wrote: »

Advantages of a hierarchical code

1. Easier to learn/remember. Structured information is always easier to remember than random information. Because hierarchical codes are more "recognisable" there is less scope for error when calling them out.

The postcode is two parts, the first part has hierarchy the second doesn't. You will remember them as two memory chunks, therefore additional hierarchy in the second part of the code will not aid memory recall. In the same manner that you don't have difficulty remembering your mobile phone number even though it is very different to your neighbours.

2. Small businesses are not likely to license the Eircode datasets. So, Eircode is of limited use to them. The typical example is a tradesman who has a list of callouts in a day and wants to organise his traveling time efficiently. With a hierarchical code, he can do that just by looking at the code. With Eircode, he will have to buy some third party product (assuming such a thing will exist) to disentangle the random codes. Why impose technology on people for its own sake (remember evoting!)

Small business have no need to license Eircode. They will use products that contain postcode data. Using hierachy in a postcode will generate a more efficient route than random ordering, but won't generate the most efficient route, which requires technology. This is known as the Travelling Salesperson problem.

3. Data protection concerns. For some purposes (eg anonymous market surveys) a hierarchical code allows people to disclose where they live, but without disclosing an exact address. Not possible with Eircode.

Solutions aren't difficult when you actually try to think of them rather than simply wave your hands in the air. Input postcode into "privacy" app, output Small Area, give to market survey person. Alternatively data protection requirements may make this standard practice on behalf of the receiving company, take postcode but don't store it, store the Small Area instead.

4. New house builds. The one time when a householder most needs an accurate post(location) code, is when a new house gets built. It's likely to take months before an Eircode is allocated, reflected in a quarterly update, and then updated in third party applications. A hierarchical code will have some level of position accuracy from before a house gets built since the area (in most cases) pre-exists the house.

You can't "predict" your postcode hierarchy accurately, even then it is only a partial postcode, you don't have a full one.

5. Better for emergency services. If a householder wants to call the emergency services to their area, because there is an incident somewhere nearby, but not at their own property they can't simply specify the area. They have to provide their own code, which could result in callout charges being directed to them.

I see this argument has moved on from the initial "with a hierarchical code you could give them your postcode, with a random one it is of no use" which was ludicrous to now you could receive a callout charge for a property that isn't yours because you've phoned the emergency services. Utter nonsense.

6. Better for government. At least two government departments said they wanted a hierarchical code. I don't think any said they wanted a random code. The advantages are obvious in terms of a consistent view of where people live. Also, small areas can be aggregated in different ways. So, the postcode doesn't impose one particular structure for all purposes. Eg you can combine small areas in one way to make up electoral divisions, and in a different way to make up school districts.

There are NO benefits for government for a hierarchical code. It is already unique to each address. This allows any boundary to be used as required. Small areas are simply one such example. Your point is completely invalid.