Scan lists require non-volatile memory. Memory costs $$$. It is always a trade-off in determining how much memory to design into a product. This is also why I would like to see something like support for an (SD) memory card slot. Then each user can use whatever amount of memory they desire (to purchase).
Actually, scan lists in the GRE/RS high end units are not actually memory locations. This is a common misconception caused by carrying the "scan list = bank" analogy too far. Scan lists, as used in the PSR500/600 and PRO-106/197, are better thought of as tags or pointers to actual memory locations. The memory locations in the scanner are represented as blocks and cataloged by the Object ID number. Objects take up one or more blocks of memory depending on the kind of object (TSYS, TGRP, conventional frequency, etc.).
This is why you can have an object belong to any number of scan lists as you like but not use up any more memory space as would be the case in a bank scanner. One object could belong to one or all twenty-two scan lists and the same amount of memory would be used either way. Each scan list also has no real "limit" in the sense of how many objects can belong to it. You could have every object in a scanner belong to one list and even if the entire memory of the scanner was used it would be no different then if no objects belonged to any list or if each belonged to different lists. This is actually very efficient and quite flexible.
Nevertheless, and perhaps even more so because of the above, I, too, would like to have more scan list categories. Due to the above, rather than total memory capacity being the limiting factor it is more likely that some form of memory structure construction issue is the problem. This might be some low level code that was written to form the overall structure of the memory (22 scan list limit, 32 frequencies limit for TSYS's. etc.) and to change it might be either impossible or very difficult (requiring a near total rewrite of the memory organization of the scanner).
Personally, I would like to see another layer of categorization in addition to the scan lists (say sub lists). One way of doing this is to create something akin to the "TSYS" for conventional frequencies as someone else mentioned some while ago and called it a "CSYS". This is because the TSYS's can be used to add a layer of organization in addition to the scan lists - great for trunking systems but of no help to conventional frequencies.
Ideally, I would like to be able to categorize with three layers regardless of type of object (talk group or conventional frequency). For me, I would use the layers as follows:
Layer 1: Location (i.e. North San Diego County).
Layer 2: Broad Usage Category (i.e. Law Enforcement, Fire, Parks, Roads, Transportation, etc.).
Layer 3: Sub-category of usage for geopolitical specifics (i.e. County, State, Federal, Multi-use, Private, etc.).
The above layers would be user defined so you could use them any way you want - I just used examples of how I would do so. Ideally, there should also be a minimum limit of 10 divisions for each layer (with 3 layers this should be more than sufficient to specifically categorize an object in a very flexible and precise manner). Nevertheless, for maximum flexibility and to appease as many potential users as possible you could make the top layer have say 50 or 100 divisions while layers 2 and 3 could have less and the user should be allowed to not use a layer at all if so desired.
And, of course, talk groups should be allowed to belong to multiple trunking sites if desired rather than having to be entered multiple times for multiple sites. Ideally, a TSYS should define a trunking system containing a name, SYSID, type (Motorola, P25, EDACS, etc.), and modulation (wide, narrow) with capacity for multiple sites which each contain a name (if applicable), site number (if applicable), and location (if desired) plus the primary and alternate control frequencies relevant to that site. Then, talk groups could be "linked" or "associated" to one or more of these sites as desired so duplication would not be required. Alternatively, if the above cannot be done, at least allow talk groups to belong to multiple TSYS's in the current configuration if at all possible.
Also, usage of the SYSID for filtering the TSYS's would be a very useful feature IF it could be made to work effectively. Alternatively, or in addition to that, it would be useful to be able to assign priorities to the trunking site's control channels (try this one first, then this, etc.) or, if the multiple sites belonging to a TSYS option could be made to work as outlined above, then assign priorities to the individual sites (try this one first, then this, etc.). In each case, higher priority sites (or control frequencies) would be checked first on each pass and given the most due consideration during signal loss due to drop outs and fading.
I've also been toying with an idea for a "DSYS" object (for want of a better term) wherein conventional frequencies used in semi-duplex split frequency simplex systems could be scanned more efficiently than they are in current configurations. For example, the well known California Highway Patrol systems. Most of the channels used throughout the state are of the above format wherein the mobiles respond to the bases on a separate frequency and are not repeated. Traditionally, when scanning these types of systems the user would have to program two sequential channels with the base and mobile frequency respectively and leave the delay for each channel off. This "sort of" works but if the mobile does not respond quickly enough or the signal is fluctuating the mobile response may be missed (even if the signal is strong enough). Since trunking systems are set up with multiple frequencies being checked I would think it a relatively simple matter to create a type of object that has two frequency slots for a base and mobile pair. When scanned, that object looks at both of those frequencies for a specific time before moving on. If either frequency has a signal with the correct tone and sufficient strength, etc., the scanner stops on that object and unmutes; when the signal stops it quickly scans between only those two frequencies within that object for a specified time (just the user defined delay time as used on any other object) before moving on. This would make it far more likely that one could catch the mobile response if within range than using the traditional method. Additionally, the two frequencies should be able to have their own respective squelch modes, CTCSS, DCS, even NAC, for maximum flexibility - I know the CHP don't (as far as I know) use separate tones but I would want the system to be as flexible as possible and not limited to just one methodology. Many taxi systems also use a similar method. Also, because when looking at this object the scanner is rapidly scanning between these two frequencies, it will not matter if the mobiles are talking to each other on the base output frequency, as the CHP routinely does for car-to-car use, as they will still be caught if in range. Oh well, just a thought.
I don't necessarily expect that all of the above would be feasible for these current units with a simple firmware change but - perhaps on future units.
-Mike