I confess to being a bit surprised about the outcome of the polling thus far. The majority appear to want 5 categories, roughly speaking: small, medium, large, huge and gargantuan. Some want just 4 categories, leaving off the small 'hatchling' category as uninteresting. Some even want 3 categories, typically dropping out either the medium or the large category as well.
Looking at my current work, and imagining cutting back to 3-5 categories, the big problem that I see is the gap in challenge gets really large.
ccs expressed this problem the best when he wrote: "However you break it down it needs to accommodate my minis - which range from small bases up to that colossal red with the removable fire breath WoTC made back in 3.5"
When you get to the tangibles of dragons, small, medium, large, huge, gargantuan, and colossal are all real things of very noticeable difference in size. Looking at my own work thus far, the low end of the range is a reasonable challenge for a 3rd level party, and a difficult challenge for a 1st level party. The high end is a reasonable challenge for a 16th level party, and a difficult challenge for a 14th level party. Fitting just one or two intermediary categories into that means very large gaps in the challenge rating. I'm not seeing that the added complexity of another category adds more problems than the gaps do, at least when the amount of information required to build and record a category is very small.
Consider the process of converting my work to 4e. The 'hatchling' category represents CR 1. The largest most legendary dragon represents CR 30. To get to 5 categories, its only necessary to put one dragon at the midway point of each tier - CR 5, CR 15, CR 25. But look at those gaps now between CR 5 and CR 15, and between CR 15 and CR 25. I'd argue that it's pretty obvious even 4e needs 7 categories to have any sort of smooth progression. To argue otherwise feels like arguing to me that PC's only need 3 or 5 levels to represent play. Why then do they have 30?
The recording of age category in 1e is very small indeed. It mostly means recording that the hit points per dice are equal to the age category, plus some special powers (like awe) that depend on having age category. The problem IMO is that this convenient means of recording doesn't record enough information (well, technically it's complete, but it results in simplifications that are bizarre). 2e tried to fix this, but in my opinion the 2e version went too far the other way. Recording age category in 2e and 3e requires a relatively large amount of information. In 4e, stat blocks are very disconnected and very large. There is a lot of cost in adding a stat block, and 4e very much deprecates the idea of flexible blocks (stat blocks that are implied or tabulated by the DM) in favor of showing all the rules together. The literal cost of adding age categories to a manual is high, as it eats a page or more. But that doesn't mean say 7 categories wouldn't be useful or that the granularity between them would be trivial.
In short, this thread has achieved for me it's primary goal - proving that on the whole, most people wouldn't object to fewer than 12 age categories. My feelings haven't really changed though, that depending on edition, 6-10 age categories have value in the context of that edition. Some in this thread seem to prefer the low end of that spectrum. I tend to lean to the high end of that spectrum. To convince me to adopt the low end of the spectrum, you'd need to convince me that objections like, "You already have a large and a gargantuan option, so you have no need for a huge dragon.", or "You already have a CR 8 and a CR 12 option, so you have no need for a CR 10 option.", are very valid objections.
The goal of my project can be thought of as this. Most monsters did not require major redesign between 1e and 3e. There are clear parallels between the stat blocks of most monsters in their 1e, 2e, and 3e versions. To a lesser extent, this is also true of the 1e and 5e stat blocks. Had 1e dragons been designed correctly in the first place, there would have been no need for 2e's large refactoring, and we'd see a clear parallel between the 1e, 2e, and 3e versions as well.