Eircode design

antoinolachtnai

You can make the case for a randomised code if you like, and that is fine. There is an obvious problem with it, but if you are happy enough paying for proprietary databases and waiting for database lookups, that is fair enough.

The logical implication if you accept that is that the whole code should be random, not just the latter part of it.

The whole thing about structuring the first part of the code around delivery areas for flat post is a load of rubbish. The flat post business is dying and will be dead in 5 years.

plodder

Bray Head wrote:

Of course, many people would feel discommoded if they felt part of a more prestigious area but were in a different postcode. But that would be a reaction to their particular postcode, not the design in general. People who felt lucky to be included in a more desirable postcode would be silent.

The random nature of the code system had its disadvantages. I won't repeat them because I think on balance the design was the right one.

One very positive outcome of the chosen design is that there is virtually no objection to it on snobbery grounds (over and above what already existed in Dublin).

The smaller the area that is visible in the code, then the less likely the problem is imo. Eircode went to one extreme, where there are no areas (below the rk), but imo "small areas" would have been fine for the most part. If you live in "Downbeat drive" with postcode C12 XXX, it's harder to claim you should have the same postcode as the people who live in "Posh place" C13 XXXX, because you don't live in Posh place. When you get to larger areas like Togher and Wilton, it gets more difficult because the boundaries are probably different for different purposes and the potential for disagreement greater.

Bray Head wrote: »

People would probably have opted for some derivative of the British system because that what they are culturally familiar with.

I would congratulate our policymakers (for once) for not choosing a British solution and instead opting for a system which is more suited to Irish needs (non-unique addresses) and a technology infrastructure that has seen 50 years of progress.

That old chestnut. Like evoting all over again. As if something is automatically obsolete because it is old... As someone pointed out on the other thread recently. You type in a random code on Domino's website to pinpoint your location here. But, you type in your postcode + housenumber in the UK to do the same. Is there really such a big difference there to justify your "50 years of progress" claim?

People who actually know this stuff would say the British system is better because it keeps the postcode as an anonymous identifier, more in keeping with other countries postcodes. Then you don't get organisations like RTE in that survey having to remove people's postcodes from it before passing the data on.

In any case, the point of this kind of research is to find out what people think rather than assuming what they will find acceptable. And funnily enough, I would often criticise us for copying the UK blindly, but I wouldn't so much in this case. And lastly, I'd say it again, these controversial aspects of the design could be partially side-stepped with a free dataset. You might be able to make a stronger case for the randomness aspects then, because you would at least be offering people the ability to decode it for free down to a certain level.

ukoda

plodder wrote: »

The smaller the area that is visible in the code, then the less likely the problem is imo. Eircode went to one extreme, where there are no areas (below the rk), but imo "small areas" would have been fine for the most part. If you live in "Downbeat drive" with postcode C12 XXX, it's harder to claim you should have the same postcode as the people who live in "Posh place" C13 XXXX, because you don't live in Posh place. When you get to larger areas like Togher and Wilton, it gets more difficult because the boundaries are probably different for different purposes and the potential for disagreement greater.

That old chestnut. Like evoting all over again. As if something is automatically obsolete because it is old... As someone pointed out on the other thread recently. You type in a random code on Domino's website to pinpoint your location here. But, you type in your postcode + housenumber in the UK to do the same. Is there really such a big difference there to justify your "50 years of progress" claim?

People who actually know this stuff would say the British system is better because it keeps the postcode as an anonymous identifier, more in keeping with other countries postcodes. Then you don't get organisations like RTE in that survey having to remove people's postcodes from it before passing the data on.

In any case, the point of this kind of research is to find out what people think rather than assuming what they will find acceptable. And funnily enough, I would often criticise us for copying the UK blindly, but I wouldn't so much in this case. And lastly, I'd say it again, these controversial aspects of the design could be partially side-stepped with a free dataset. You might be able to make a stronger case for the randomness aspects then, because you would at least be offering people the ability to decode it for free down to a certain level.

The UK system of postcode plus house number would be entirely useless for over 30% of Irish addresses.

Also, in a change of pace, I agree with you that it would be a good idea to release a dataset for free.

plodder

ukoda wrote: »

The UK system of postcode plus house number would be entirely useless for over 30% of Irish addresses.

You could assign a house number where none exists. It would mostly be a two digit number. But, we are where we are, and it isn't this way. So no point dwelling on it (pun not intended)

Also, in a change of pace, I agree with you that it would be a good idea to release a dataset for free.

Wonders will never cease :pac:

Bray Head

antoinolachtnai wrote: »

You can make the case for a randomised code if you like, and that is fine. There is an obvious problem with it, but if you are happy enough paying for proprietary databases and waiting for database lookups, that is fair enough.

The logical implication if you accept that is that the whole code should be random, not just the latter part of it.

This is a fair point.

If I was starting from scratch I would advise a totally random code. But once the decision to use the Dublin postcodes was taken it was kind of inevitable that the rest of the country would be on a geographical basis too.

Sam Russell

Bray Head wrote: »

This is a fair point.

If I was starting from scratch I would advise a totally random code. But once the decision to use the Dublin postcodes was taken it was kind of inevitable that the rest of the country would be on a geographical basis too.

A purely random code becomes a PPS for addresses - not sure that has any merit without free access to the database and not even then.

The current design falls down on the adherence to the An Post routing codes which are basically temporary (in the scheme of things) and anyway are pointless. Because they still need a computer driven scanning and sorting system then why adhere to current systems that can and will change even in the short term (next ten years).

Until people nail their Eircode to their gatepost, then knowing an Eircode does not work for the casual user.

Until everyone puts their Eircode into the public domain, it is hidden from most potential users.

Maybe they (Eircode) should work on those aspects.

ukoda

Sam Russell wrote: »

Until people nail their Eircode to their gatepost, then knowing an Eircode does not work for the casual user

Unless they have a smartphone. Which is pretty much everyone.

Bray Head

A fairly innocuous story on the journal about two mountain walkers being rescued turned into a lengthy paean to the benefits of the Loc8 system

Eircode also gets a mention: http://www.thejournal.ie/mayo-mountain-rescue-2913227-Aug2016/

gctest50

Bray Head wrote: »

A fairly innocuous story on the journal about two mountain walkers being rescued turned into a lengthy paean to the benefits of the Loc8 system

Eircode also gets a mention: http://www.thejournal.ie/mayo-mountain-rescue-2913227-Aug2016/

This nugget of genius from the GetLostEircodes clown :

For recognised walking routes way marked trail posts could contain a written position using Loc8 code every 200m…no GPS or data signal required. - GetLostEircodes

No mention of http://what3words.com

sondagefaux

Some issues with UK postcodes identified in the comments section here:

http://jonathan.rawle.org/2006/07/03/postcodes-in-the-uk/#comments

plodder

PDVerse wrote: »

Your point is simply invalid. There are opinions, informed opinions, and expert opinions. To give an example:

Expert Opinion: Climate change is real
Informed Opinion: I'm not a climatologist, but I've studied the statistical data and I have questions.
Opinion: See Danny Healy-Rae

CSO's job is to gather and publish statistics in various formats suitable for their use case. As someone who has worked with their output for 23 years I can assure you they do a very good job. Small Areas were created specifically for CSO to enable more granular statistics below ED level. They DO NOT have a deficiency that they were hoping a postcode design would resolve. You're trying to create a problem that doesn't exist.

The road from ignorance to knowledge requires honesty. You have to admit that you only have an opinion, and to legitimately argue against expert opinion you have to at least put the work in to be informed.

Appeal to authority. "I designed Eircode. Therefore, I must be right"

It's quite a simple point really and maybe you might confine your reply to it rather than giving your views on climate change denial etc.

How useful is a house price statistic for the H91 area, when it covers a population of around a quarter of a million people, when there are routing key areas in the Dublin area with populations in the low thousands? It's a fact that you can't distinguish trends from the different areas within it, like Galway city, East Galway, Conemara etc. Simple question.

PDVerse

plodder wrote: »

Appeal to authority. "I designed Eircode. Therefore, I must be right"

It's quite a simple point really and maybe you might confine your reply to it rather than giving your views on climate change denial etc.

How useful is a house price statistic for the H91 area, when it covers a population of around a quarter of a million people, when there are routing key areas in the Dublin area with populations in the low thousands? It's a fact that you can't distinguish trends from the different areas within it, like Galway city, East Galway, Conemara etc. Simple question.

Argument from authority is an invalid argument. However it doesn't follow that everyone's opinion thus holds equal weight which is what I was explaining with my analogy.

Thank you very much for pointing out that the design of Eircode does not easily allow house price discrimination. Mission accomplished.

The Routing Key areas were solely defined by An Post, to suit their requirements. When something is designed for one purpose you generally find it does a poor job when applied to a completely different purpose.

There was no deficit in publishing statistics that required a solution from a postcode design. This is not a debatable point, go talk to the CSO or someone other than me that also uses their data on a daily basis.

plodder

PDVerse wrote: »

Argument from authority is an invalid argument. However it doesn't follow that everyone's opinion thus holds equal weight which is what I was explaining with my analogy.

Thank you very much for pointing out that the design of Eircode does not easily allow house price discrimination. Mission accomplished.

The Routing Key areas were solely defined by An Post, to suit their requirements. When something is designed for one purpose you generally find it does a poor job when applied to a completely different purpose.

Thank you for acknowledging that. More proof that An Post shouldn't have had such a central role in the design.

There was no deficit in publishing statistics that required a solution from a postcode design. This is not a debatable point, go talk to the CSO or someone other than me that also uses their data on a daily basis.

I wouldn't agree with that. Postcodes have enormous potential for statistical use. Otherwise, the CSO wouldn't be using them now.

PDVerse

plodder wrote: »

Thank you for acknowledging that. More proof that An Post shouldn't have had such a central role in the design.

I wouldn't agree with that. Postcodes have enormous potential for statistical use. Otherwise, the CSO wouldn't be using them now.

In summary you've imagined an issue that CSO doesn't actually have, and want to complain that the Postcode design doesn't solve it.

I'm happy to discuss other aspects of Eircode design with you, but I'm afraid I've reached the limit of my patience with this particular topic.

plodder

PDVerse wrote: »

In summary you've imagined an issue that CSO doesn't actually have, and want to complain that the Postcode design doesn't solve it.

Hmm. I never said the CSO had an issue. The issue is with Eircode. The CSO will just publish their stats as normal.

I'm happy to discuss other aspects of Eircode design with you, but I'm afraid I've reached the limit of my patience with this particular topic.

Part of the reason why I'm doing this, is to (maybe vainly) hope that the "powers that be" realise, for future work, it's far better to have these open, public discussions before these projects get designed and delivered, rather than afterwards, when it's too late and they take on the excruciating character we see here, where some people are incapable of conceding the tiniest criticism.

Sam Russell

I have never understood why Eircode was designed to facilitate An Post when they were of the opinion that they did not need it and they had a fully automated sorting system. It is like building the motorways to suit Irish Rail.

If they had used 1,000 or more routing codes instead of 139, it would not affect them one jot but would have increased the usability of Eircode for many users.

Using routing codes that cover 250,000 people is just nuts in our population. Why bother at all - a more reasonable maximum number would be 25,000 - a small to medium town.

PDVerse

Sam Russell wrote: »

I have never understood why Eircode was designed to facilitate An Post when they were of the opinion that they did not need it and they had a fully automated sorting system. It is like building the motorways to suit Irish Rail.

If they had used 1,000 or more routing codes instead of 139, it would not affect them one jot but would have increased the usability of Eircode for many users.

Using routing codes that cover 250,000 people is just nuts in our population. Why bother at all - a more reasonable maximum number would be 25,000 - a small to medium town.

An Post do not have a fully automated sorting system. No one does. Some post cannot be read automatically, and must be sorted manually. Unfortunately, busy periods like Christmas and Valentines correspond to increases in automation recognition issues caused by feint handwriting on brightly coloured envelopes. It is imperative that the first part of the postcode allows An Post to efficiently sort manually.
An Post chose 139 principal post towns to ensure efficiency (different mail centres sort into different groups) and be future proof. By definition they are fit for purpose.
The more Eircode is used on post, the more An Post will use Eircode.

As I outlined in a previous post, the size and number of Routing Keys is a non-issue for implementations. The idea that this design is restricting the usability of Eircode for many users is a hypothesis posited on this thread that is at odds with reality.

I'm looking forward to Google API integration of Eircodes so it can be used in various ways, usually entirely free. It will also demonstrate that the design wasn't influenced by a desire to "charge for even basic usage".

plodder

Sam Russell wrote: »

I have never understood why Eircode was designed to facilitate An Post when they were of the opinion that they did not need it and they had a fully automated sorting system. It is like building the motorways to suit Irish Rail.

Good analogy.

If they had used 1,000 or more routing codes instead of 139, it would not affect them one jot but would have increased the usability of Eircode for many users.

Considering they had to update the automated scanning equipment anyway to recognise the routing keys, I can't understand why they didn't split some of the larger areas into a number of smaller ones.

Using routing codes that cover 250,000 people is just nuts in our population. Why bother at all - a more reasonable maximum number would be 25,000 - a small to medium town.

Exactly, while it's not a problem for many applications that basically don't care about routing keys, there are use cases where it does matter, as the CSO stats show.

PDVerse

Fictional Headline in The Newspaper dated April 2014

Ireland's new Postcode is not a Postcode say An Post.

Our reporter has learned that the new postcode design document has been met with disbelief by An Post. "We gave very clear requirements how the Routing Key needed to be structured to suit mail delivery. We spent a year working on this, and its been totally ignored. They claim they needed to accommodate CSO reporting. Whatever else this code is, it isn't a postcode."

Our reporter contacted CSO who expressed bewilderment that the design that suited An Post had been ditched in order to accommodate their requirements. "We're at a loss to explain that decision. We explained that we had no requirement for a hierarchy in the postcode design, once it was unique to each address it suited all our requirements. We already have a hierarchy that has been specifically designed to meet our requirements, why would we require another hierarchy in the postcode to do a similar job, but less fit for purpose?".

A spokesperson for the design team was adamant the right decision had been made. "We had open, public discussions and there was overwhelming support from the public we surveyed for a postcode that would assist CSO to provide better statistics. The spokesperson refuted our reporters suggestion that they had ignored the requirements of An Post in order to solve a problem that CSO say doesn’t exist.

Sources within the Department say they are livid with the design team for failing at their fundamental duty of designing a postcode that can be used as a postcode and now regret not awarding the tender to An Post.

A source within the Cabinet claims the project may be quietly dropped over the summer.

plodder

So, if An Post produced the current list of routing key areas and Eircode said. That's fine - except we'd like to split some of the larger ones into smaller areas, to make them more consistent, so you'll have to program those scanners to recognise 4 or 5 different routing keys instead of one for each of them.

If Eircode did that, you're saying they would have cried foul and said "This is no longer a postcode"?

That's before you even consider the fact that An Post made it clear they never needed a postcode, and you wonder who actually runs the country.

gctest50

PDVerse wrote: »

Headline in The Newspaper dated April 2014

Irelands new Postcode is not a Postcode say AnPost.

Our reporter has learned that the new postcode design document has been met with disbelief by An Post. "We gave very clear requirements how the Routing Key needed to be structured to suit mail delivery. We spent a year working on this, and its been totally ignored. They claim they needed to accommodate CSO reporting. Whatever else this code is, it isn't a postcode."

Our reporter contacted CSO who expressed bewilderment that the design that suited An Post had been ditched in order to accommodate their requirements. "We're at a loss to explain that decision. We explained that we had no requirement for a hierarchy in the postcode design, once it was unique to each address it suited all our requirements. We already have a hierarchy that has been specifically designed to meet our requirements, why would we want something less fit for purpose?".

A spokesperson for the design team was adamant the right decision had been made. "We had open, public discussions and there was overwhelming support from the public we surveyed for a postcode that would assist CSO to provide better statistics. The spokesperson refuted our reporters suggestion that they had ignored the requirements of An Post in order to solve a problem that CSO say doesn’t exist.

Sources within the Department say they are livid with the design team for failing at their fundamental duty of designing a postcode that can be used as a postcode and now regret not awarding the tender to An Post.

A source within the Cabinet claims the project may be quietly dropped over the summer.

July 14, 2016 :

July 14, 2016 : An Post said Eircodes have been fully integrated into its automated sorting systems and it "fully supported and continues to support the introduction of the new Eircode system".

https://www.rte.ie/news/2016/0714/802305-eircode/

PDVerse

plodder wrote: »

So, if An Post produced the current list of routing key areas and Eircode said. That's fine - except we'd like to split some of the larger ones into smaller areas, to make them more consistent, so you'll have to program those scanners to recognise 4 or 5 different routing keys instead of one for each of them.

If Eircode did that, you're saying they would have cried foul and said "This is no longer a postcode"?

That's before you even consider the fact that An Post made it clear they never needed a postcode, and you wonder who actually runs the country.

You seem to continually miss the requirement: Manual sorting of post.
An automated scanner can read anything. Some of the manual processing is simply keying the correct code when displayed on a monitor, but other processes involve physical sorting into slots/bags. Any changes to these existing processes would cause extensive disruption that would require real rather than imagined benefits to justify.

You don't design a postcode, then trundle up to An Post and ask how much of an inconvenience would it be for them to change their current and future manual sorting to accommodate it. You start with their requirements. The pros and cons of splitting Routing Keys is covered in the design document.

You seem to be very upset that the design team ensured that the postcode could be used as a postcode, but I honestly can't think of any reason to justify your position.

plodder

PDVerse wrote: »

You seem to continually miss the requirement: Manual sorting of post.
An automated scanner can read anything. Some of the manual processing is simply keying the correct code when displayed on a monitor, but other processes involve physical sorting into slots/bags. Any changes to these existing processes would cause extensive disruption that would require real rather than imagined benefits to justify.

You don't design a postcode, then trundle up to An Post and ask how much of an inconvenience would it be for them to change their current and future manual sorting to accommodate it. You start with their requirements. The pros and cons of splitting Routing Keys is covered in the design document.

You seem to be very upset that the design team ensured that the postcode could be used as a postcode, but I honestly can't think of any reason to justify your position.

I'm not actually upset at all, but it's worth pointing out that An Post originally didn't want a postcode. They invested very heavily in automated technology and originally said that postcodes were a "1950's solution" to the problem. So, I'm sceptical.

They can sort manually or automatically perfectly well with existing addresses, minus Eircodes. In reality, what Eircode adds to the postal system, is disambiguation of non-unique addresses. Though allegedly An Post weren't too keen on that either as it was bypassing the local knowledge of their own staff.

And none of the above adequately explains why they would have to change any manual process, had the larger routing keys been divided into a few smaller ones. As you say scanners can be programmed to do anything. For manual sorting (by Eircode) you would just have 3-4 alternative routing keys as opposed to just one in some cases. Presumably, for manual sorting they have a list of routing keys mapped to internal sorting slots/bags. There's no way that any manual process has to change in that scenario.

Instead of

Mail for H91 ---> Slot A
Mail for F31 ----> Slot B

you might have

Mail for H90 ---> Slot A
Mail for H91 ---> Slot A
Mail for H92 ---> Slot A
Mail for F31 ----> Slot B

where H91 today is H90,H91 and H92 instead.

I think the real reason, is that a design decision was made to de-emphasize areas in the public part of the code completely. Areas were seen as trouble basically. But as I predicted, people (eg the CSO) will still use them regardless.

Curly Judge

Why has my postal address been changed, arbitrarily, to suit Eircode?

oscarBravo

Curly Judge wrote: »

Why has my postal address been changed, arbitrarily, to suit Eircode?

It hasn't.

Right2Write

All I know is that very few people seem to know or use Eircodes. Why bother seems to be the attitude, when the system wasn't broke in the first place?

ukoda

Right2Write wrote: »

All I know is that very few people seem to know or use Eircodes. Why bother seems to be the attitude, when the system wasn't broke in the first place?

Wasn't broke??? What part of 30% of the countries addresses being impossible to pinpoint was ok?

Curly Judge

oscarBravo wrote: »

It hasn't.

It has!
If I use my old address....the one I and my neighbours have been using for all our lives....and add the new eircode alphanumeric at the end....it will be rejected by credit card companies and, in the case of mail, one will get a snappy note from An Post telling me to to use the "proper" address.
In the case of mail, it only seems to happen if the automatic sorter kicks out the letter because of an unreadable address and it has to be reviewed by a human operative.

TheChizler

Curly Judge wrote: »

oscarBravo wrote: »

It hasn't.

It has!
If I use my old address....the one I and my neighbours have been using for all our lives....and add the new eircode alphanumeric at the end....it will be rejected by credit card companies and, in the case of mail, one will get a snappy note from An Post telling me to to use the "proper" address.
In the case of mail, it only seems to happen if the automatic sorter kicks out the letter because of an unreadable address and it has to be reviewed by a human operative.

That's those companies rejecting it, not Eircode. They've chosen to compare the Eircode to the postal address only instead of looking at both.

Curly Judge

TheChizler wrote: »

That's those companies rejecting it, not Eircode. They've chosen to compare the Eircode to the postal address only instead of looking at both.

But why is Eircode coupled, paired, and associated with a new.... never before used ......address?