As the Government continues to struggle with the implementation of an Irish postcode the issue is very confused and the awful part is that those in Government circles may never realise in time how simple the coding and how much more accurate and useful an opensource postcode can be. Governments get drawn in by talk of centralised databases, postcode management, and claims of accuracy which do not exist - while all along it is no more difficult than a simple conversion of a geographic coordinate - coordinates which the government already own in the GeoDirectory. Similarly there is no need for central management of postcodes and no urgent need to have every single address "postcoded" in one fell swoop - various databases, such as household charges, landregistry, etc., will develop through use as with anything in the open world.
In mathematical terms, to convert coordinates into the OpenPostcode (accurate to an area 5.98m x 4.6m in seven characters), for example for the Spire of Dublin at 53.349795, -6.260254 is a simple as:
base25to10(base10to5(int((55.5-53.349795)/4.2*390625))))*5 + base25to10(base10to5(int((10.75-6.260254)/5.4*390625)))) = "EAIONL59"; which makes "KFQZXV7D" by swapping the normal base 25 character set for the OpenPostcode set of "23456789CDFGHJKLMNPQRTVWX".
Millions complete in the blink of a computer's eye.
- The Government proposed postcode model of “ABC-123” to be centred about 200 English language “post towns” is incapable of providing any better than a generously estimated average 0.49km² precision - which does not meet the DCENR’s own specification of a postcode identifying a group of 40 to 50 addresses. Two addresses in the same postcode could be nearly 1 km apart as the crow flies - bigger than most villages and large enough to encompass the heart of many small towns. In rural areas it is a bit better - not quite as bad as the traditional vague townland address but still not accurate enough to overcome the rural addressing problem in Ireland without the addition of house numbers. However, it copies an unwieldy, outmoded, unsuitable UK postcode system. The proposed code will provide less location information than the ordinary address. Further analysis of the proposed Government model is given here.
- GeoDirectory is an autonomous company under An Post with a location database of 1.7 million addresses geocoded for their use. This code be converted to the OpenPostcode immediately for free.
- Loc8code Ltd. have an eight character "Loc8code" licensed coding calculated from OSI Mapping (ITM based rather than GPS based) only which is claimed to be able to "define locations to an accuracy of +/-6 metres". The code also claims a 0m offset "repeatability" although it is only smoke and mirrors - a database remembers a single mappoint chosen when creating a code, the code itself doesn't have the level of accuracy claimed and the trick cannot be repeated offline or on portable devices without access to the proprietory database and is impossible with the many locations that have to share the code's 110.27m2 (measured from loc8code field tests). The code's last character is a "checker" but has little or no effect and only seems to check itself and be an entire waste of data. The Loc8code has a mixed character set of letters and numbers, avoiding some for clarity and to avoid accidental language; only starts with a letter; while it uses just plain numbers for the middle two digits. While using a grid to define the island the code maximises precision by off-setting the grid to match the slight angle of the island (as described by a loc8code Ireland map). This interrupts a direct calculation between the code and location. Loc8Codes do not sort correctly (though since they can't be freely converted to coordinates not really relevant anyway). The Loc8code has limited scaleability: offering three levels of precision. There is no open description of the character or calculation of the code. It is not free to use for calculations. It cannot work as a geolocation code without recourse to its proprietary calculations. It will add no geolocation analytics to the smart economy - it can only work as an arbitrary jumble. Top marks for promotion and for an extensive website but it's just packaging for a below par, confused, code.
- Go Code Ltd. propose a seven character "GoCode" stated to be "up to an accuracy of 5 square metres approx". It is the shortest basic code. However it does not include a checksum. The code uses a 25 character set, except for the first character which is a 29 character alphabet (including 0, 1, "e" and "y"). There is some very minor accidental language caused by this - nothing dramatic. Using 29 characters for the first character increases accuracy significantly. Furthermore, the code does not draw a complete orderly square about the island. This interrupts a direct calculation between the code and location. Codes sort correctly to some extent (though since they can't be freely converted to coordinates not really). The code is scaleable and conceivably extensible - it would have to change slightly however for other territories. It does not have a published calculation. It is not free to use for calculations. It cannot work as a geolocation code without recourse to its proprietary calculations.