Yesterday, Pat Donnelly of Autoaddress, a company which is part of the Capita consortium that designed and delivered Eircode, wrote in a blog post that it was not possible to have included useful localities in the Eircode design.
Eircode has no localities and this is contrary to the legal requirement for a National postcode as specified in the Postal Act of 2011, the specification of which resulted from the recommendations of the Government appointed National Postcode Board in their report of 2006. That report required the National Postcode to be structured to “include small spatial areas” and that formed the basis for the tender from which Eircode emerged.
In these regards, the legality of Eircode as the National postcode, which emerged from a procurement process that sought something very different, is seriously in question.
Mr. Donnelly makes the argument that including such localities would have been difficult and that is indeed agreed. However, Mr Donnelly and his fellow consortium members were paid €2 million of taxpayers’ money to come up with a workable design which recognised the requirements and overcame the difficulties in achieving that, but his recent blog post essentially admits that they failed to achieve this and, therefore, have not delivered value for money. Whilst the design of Eircode may have cost €2 million, the cost to the taxpayer of delivering it as the National Postcode now stands at €50 million approximately.
Mr. Donnelly also fails to admit that in a period of less than 1 year after rolling out Eircode, his company had already started to promote another code as an add-on to Eircode. This new code has his own company’s IP and contains a rational answer to the same “locality” issue which he says could not be achieved in the Eircode design for which his company was paid. Though not perfect either, these newly Autoaddress coded localities are also hierarchical and have some rational order and sequence; – again desirable characteristics which he & his consortium said they were also unable to include in the €2 million Eircode design.
In June 2015, I appeared in front of an Oireachtas Committee to highlight the major legal and substandard design issues associated with Eircode. I also highlighted that Loc8 Code had recognised the related design challenges identified by Mr. Donnelly but had dealt with them in a way which still delivered usable and useful localities and structure in the code. This all was offered to the State at no cost to the taxpayer but this offer, made on several occasions, was never met with a response.
I also suggested that there was likely to be other codes emerging to address the glaring design issues associated with Eircode and since then Mr. Donnelly has proven this correct.
A report by the Comptroller & Auditor General in Sept 2015, and a related debate at the Public Accounts Committee in Jan 2016, highlighted that the Eircode procurement process suffered from a serious lack of oversight and compliance with good practice and that Eircode was unlikely to achieve any of the benefits intended for it. Mr. Donnelly’s major design failures serve to prove those regrettable conclusions.
The content of my report to the Oireachtas is on the Oireachtas Website here and these are the relevant extracts which show the content of Mr. Donnelly’s latest blog to be technically and legally unsound.
A PRESENTATION ON EIRCODE & THE NATIONAL POSTCODE – LEINSTER HOUSE 30 June 2015
5. Legitimacy of Eircode
Eircode has questionable legitimacy for the following reasons:
a. It fails to satisfy many, if not all, of the requirements of the intended National Postcode as set out by the National Postcode Project Board (NPPB) in 2006. These recommendations were arrived at after wide and documented consultation by the contracted consultants including with my own Company GPS Ireland in 2005. The full report in which these recommendations are made is on the DCENR website here. Very specifically, a postcode which is structured to at least small spatial areas (including splitting up postal areas of Cork, Limerick Cities and others) and which is also postal operator neutral as recommended has not been delivered. The full recommendations of the NPPB in its 2006 report officially formed the basis of the procurement process started in Jan 2011 and Eircode differs greatly from the majority of these recommendations. It is normal where the final procured product specification is to differ greatly from that specification laid down in the official tender PQQ, that the tender would be recommenced.
b. In multiple reports during the period 2006-2010 by the contracted consultants, PA Consultants, and also published on the DCENR website, the ABC 123 postcode model was recommended for adoption. This model was also stated as that required in the official tender specification. It was the subject of much media discussion between 2008 and 2011 and I personally have a letter from former Minister Eamonn Ryan TD in which he insists that this was the recommended model and that he had no plans to change it. PA Consultants wrote a letter to the Sec. Gen DCENR in Oct 2010 insisting that the proposed ABC 123 model be adopted as it would support both inward and outward manual sorting of mail, something which they identified as essential for a postcode to be a postcode. However, inexplicably Eircode greatly differs from this model in both the “ABC” and the “123” element. The ABC element of the proposed design was to correspond to 219 clearly defined An Post post town areas, however the equivalent first 3 characters of Eircode (the Routing key) deliberately does not define any areas at all and is related to only 139 “Principal” post towns thereby leading to a situation, as previously mentioned, that up to 80,000 properties in Limerick will have exactly the same routing key which does not reflect “at least small spatial areas” as recommended by the NPPB. It also means that the routing key on its own would be difficult to use for manual sorting as it does not immediately match the approximate 200 post towns that An Post currently has. In this case, the post town reference that An Post requires in the postal address would also be required to manually sort to that level, explaining why An Post is insisting, and will continue to insist, on the postal address being used and why using Eircode would present additional manual sorting challenges rather than minimising them. It is also one of the reasons why An Post is unlikely to make much use of Eircode for manual sorting;- contrary to PA Consulting’s specification. Of course, it is obvious that no human could manually sort using the last 4 characters of Eircode whose visual randomness can only be resolved by a database via computer technology. Of course, the full Eircode could be used for an automated sort but currently it is impractical for An Post to sort to better than Delivery Office level at its Hubs and only the Routing key is needed for that. There is no automation at Delivery Office level so that the last 4 random characters of Eircode, to identify the postman’s walk and the final destination address, can’t be used for manual sorting there either. Eircode does not, therefore, satisfy any of the specified or recommended features. Once again, it is normal where the final procured product specification is to differ greatly from that specification laid down in an official tender PQQ, that the tender would be recommenced.
d. Very significantly, in Paragraph 66 (1) of the same Act, which is available here on line here, the National Postcode System is defined as follows: ““postcode” means a code consisting of numbers or other characters or both numbers and other characters that identifies the locality of an address and, where appropriate, the geographic location of an address” This definition specifically requires that the selected postcode identifies a locality. This is in keeping with the Universal Postal Union’s definition of a postcode which states: “a postcode is a ‘unique, universal identifier that unambiguously identifies the addressee’s locality and assists in the transmission and sorting of mail items.” The NPPB’s technical design document for a National postcode specifically refers to localities as a key element of the recommended postcode. The document is available online here. In the National Postcode Design Report v4.0, prepared by Capita in collaboration with An Post and completed in May 2014, under Paragraph 1.4.6 Headed “Design Principles” and referring to the Routing Key it states: “For all other areas [thereby excluding existing Dublin Postal areas just discussed] the Routing Key should not refer to or be associated with the geographic place name or locality in the Irish or English language”. Throughout the document, much emphasis is placed on the fact that the Routing key will not identify any county, town or any form of defined area. Therefore, as it is clear that Eircode does not identify a locality as required by the legislation, it will not be a legitimate postcode as currently defined. Furthermore, as An Post’s existing post town system has High Court sanction and more closely conforms to the postcode definition (it defines a locality and uses the required characters for post town name and Dublin postal areas Dublin 1-Dublin 24), it can be argued that it alone has true legitimacy as a postcode under the Act. e. Therefore, if Eircode is implemented in its current manifestation, it appears that Ireland will have two postcodes in operation, one operated by An Post which has High Court confirmed legitimacy, and the proposed Eircode with extremely questionable legitimacy and which could be legally challenged with apparent ease. The State may, therefore, have great difficulty in insisting on Eircode being used as a property identifier in interactions with its agencies into the future and there also may well be significant implications for the amendments to the Act in the Bill related to Privacy and Data Protection aspects of a postcode currently making its way through the Houses of the Oireachtas.
6. Further Eircode Design Considerations
g. The deliberate exclusion of localities from Eircode, contrary to legislative requirements, and the resulting limitations of the code have already been covered. In the NPPB recommendations of 2006 referenced earlier, this requirement is referred to as the code being “structured to at least small spatial areas” However, one of the arguments used against implementing this has been the idea of introducing “postcode ghettos”. Whilst the term used exaggerates the nature of the issue being raised, the general concept is understood. This concept occurs when areas are notionally and manually grouped, normally on the basis of a postal operation. The choice to not provide useful localities in the code at all because of this is not justified under any circumstances. The solution is an easy one. It is to not make the defined areas notional or operational but instead to achieve them in a purely geographically mathematical way. In this way, they cannot be argued or negotiated. As a result, property groups such as streets and estates may well be split by the divisions but this enhances the deliberate mathematical quality, and it removes the human tendency to try establish associations by structural similarity and any possibility for change lobbying. This is achieved successfully in Loc8 Code and the resulting localities are small enough to be useful on their own and also by aggregation. Eircode’s approach to the solution makes the code less than useful. Furthermore, as routing areas are optionally and operationally expansive with no defined boundaries or extents (no geographic centre or limiting polygon) and applied, not by area association but, by which postman will serve an individual house, it opens the possibility that a Routing Key can be bought. It also means that a routing key cannot be anticipated in advance of allocation of a full Eircode for a new or temporary build and it lends itself to the idea that routing keys can be negotiated to appropriate the value of newly zoned green field sites. It is easy to conceive a situation where properties and business owners at the external limit of the proposed Limerick 80,000 property Routing Key would seek to be included so as to ensure their address appears to be associated with Limerick City rather than county;- something that can be especially valuable for a business. Once again, I believe Eircode has taken the wrong approach to its design and thereby contributed to the less than useless nature of the code.
b. Related to Letting Eircode Proceed
i. The Comptroller and Auditor General is currently investigating Eircode and has stated to the Public Accounts Committee that his office will decide whether to report or not in September 2015. It is reasonable to suggest that Eircode should not proceed until this decision is made
ix. Misinformation about the capability of Eircode must be countered so that no further complacency in relation to the management of National Property addressing is encouraged thereby causing an increased potential for loss of life in emergency scenarios.
x. A National Address Agency must be established to separately fix property addressing and related standards and to ensure that new addressing for the 1 million new properties and buildings projected over the next 30 years is to the highest international standards with public safety as the number 1 concern.
xi. The Geodirectory and related ECAD/ECAF databases should all then be combined and entrusted to the National Address Agency for the future.
xii. All IP related to Eircode must be ensured to be in the ownership of the State. It is understood that the trademark for Eircode is currently owned by Capita.
xiii. If Eircode does not apply to many areas, there is a need for the State to take responsibility for another additional coding system to address the needs of these areas. If not, there will be multiple codes in circulation with related confusion to the point that none will be useful. It should also be ensured that those who have in any way contributed to Eircode are not involved in developing, promoting or delivering any competing coding systems.