The county definitions used by the LDS in the IGI generally appear to be "ancient counties" - in other words broadly those that held sway until 1889, the date when the County of London was created. There are exceptions to this however, and it may be that they used whatever definition was used by the record office from where the data was extracted.
On the other hand, their 'London' covers parishes in both Middlesex and the City of London. Areas south of the river are generally assigned correctly to the ancient counties of Surrey or Kent, but there are exceptions to this - e.g. most Lewisham and Rotherhithe batches are listed as "Kent" and "Surrey" respectively, but other batches for these parishes are described as "London".
The LDS don't generally append 'shire' in the names of shire counties, which can make it difficult to figure out whether they mean - say, 'York' may mean the city of York, but it may mean Yorkshire.
These listings use a pre-1889 definition so that all the batches relating to one parish fall consistently within a single county. In fact, since there were various minor county boundary changes from c.1832 onwards, the definition used is for c.1850, the same as found in 'The Phillimores Atlas and Index of Parish Registers'. Note that even Phillimores isn't entirely consistent - it claims to reflect the position at c.1832 but is incorrect in respect of Durham and Northumberland.
Within each county table the sequence is alphabetical by place and generally follows Phillimores. I have tried to group batches consistently under place name headings, in many cases this has meant trying to standardise the spellings. There are many vague and misspelt place descriptions in the IGI, and this exercise is not yet complete. Some place descriptions will remain difficult to resolve as they contain too little information, (e.g. 'St Mary, Suffolk') or seem to contain multiple place or county names. Parishes created in Victorian times usually have their own headings in the table, though it can be difficult to be consistent on this point.
Churches and Hamlets
Placename descriptions in the IGI generally follow the following hierarchy - in the case of a town parish:
In the case of a rural parish, it's generally just:
There are numerous errors of attribution in the IGI, some of which can be explained by boundary changes. Others however are less easy to explain. Here are a few examples:
On this website the placename descriptions have been tidied up where possible, though some are so garbled that a precise attribution remains difficult. Typical of this is where a church dedication is unqualified by a town/city name, making any attribution ambiguous.
Under each main placename heading the sequence is firstly those batches where the place name is unqualified, followed by those where there is a second level of description - such as a church dedication, chapel name or the name of a hamlet or road etc. You may need to look in more than one place in the alphabet to find a given place, especially for those urban areas like London where the parish geography is complex. Also, a chapelry may appear under its own name, or under the name of the mother parish.