There are, indeed, a huge number of options.

Starting in the South, I've mentioned / talked about Westbury and Sailsbury here and I mentioned Weymouth and Southampton - the current stated destination of the route (but not all trains get that far now) as other alternatives.

Westbury - with good connections - has the attraction of keeping the line fairly selfcontained and should aid timtabling issues if we want a regular service at xx minutes after the hour.   Should also help reliability, especially if the service is frequent enough for a missed connection to result in "only" a delay of an hour or two and not half a day!

BUT ... Westbury flies somewhat in the face of some of the traffic analysis that's been done.  A great deal of research has been done on the Westbury to Salisbury section, and it's come down strongly in favour of two trains an hour on that section.   To stop "The Melksham Train" at Westbury would be almost like saying we don't believe that report as it would mean that the more frequesnt service just to the North of Westbury was actually thinned out before it got onto the section in question.  So ....

Salisbury looks to be a very attractive option to the South.  Traffic over the Salisbury to Westbury section would include Salisbury to Wilton (proposed park and ride) and to Warminster, Westbury and Trowbridge - linking our county's city (it is the only city in Wiltshire, isn't it?) to our county town.   Warminster shouldn't be sold short, either;  their's a great deal of military traffic there, for example - it's no co-incidence that this is the only full station that's survived past hard times all the way from Westbury to Salisbury.

It's not just local traffic on that extra leg that should be considered either.  There's a potential huge flow between Salsibury and Swindon - the City and the largest town in the county, both hives of industry, both major centres, both with major medical facilities, both with multiple ongoing connections radiating outwards like a star.

Options Penzance to Taunton, and branches between;  I find myself thinking "but they're already served via Bristol to Chippenham and Swindon" and I don't think the extra traffic from those areas through to Trowbridge and Melksham is really all that viable.  It also flies in the face of the Salisbury option in much the way that the Westbury option did, without having the merit of keeping the line local enough to be able to run an xx minutes after the hour service.

Looking North

Swindon Seems logical.   Lots of commuters up from Trowbridge / Melksham / Westbury, people travelling up there from Warminster and Salisbury, onward connections, etc.   Increased platform capacity in recent years.  Also the possibility of park and ride from Wootton Bassett. My only doubts about going at least as far as Swindon relate to (i) track capacity from Wootton Bassett in to Swindon and (ii) tieing up rolling stock in supplementing the half hourly Chippenham to Swindon Express when it could be better utilised southward.

If not Swindon, then Chippenham as the Northern destination.   With a fast, frequent ongoing connection to Swindon and London there might be a huge sense in turning around the train from Salisbury or Westbury at Chippenham;  a level connection into the "125", with the Melksham train using the down track between the half-hourly Bristol service, would make sense.   And (by way of example) a single train allocated to Westbury to Chippenham could give an hourly service ... as against a service every 2 hours if it went to Swindon; under this scenario, I think I prefer an hourly service with a change to a train every 2 hours but straght through.

Oxford. As you say, I've mentioned this separately ... because it IS an excellernt option.  Not only is there a good case all the way along the line from Salisbury to Oxford for relatively local traffic, but there's also a superb base for through traffic - tourism especially, and we could well be working with various tourism groups to promote this. 

Options Gloucester to Hereford.  I don't know.  I'm not aware of too much through traffic, and I'm very much aware that it's a shortish drive but a massive railway "doglet" from Chippenham to Kemble via Swindon.   I almost "fear" this option, frankly;  I can see an easy option for a bidder looking to meet the current requirement to run a Westbury to Swindon service very early in the morning peak hour, then send the train on to Gloucseter and use it to provide a two-hourly stoppen up and down there all day.  In the evening, it would provide a commuter service out of Gloucester and then a very late in the peak hour service on to Westbury.   Minimum service levels met, but at the price of the service through Melksham being scheduled such that it's even earlier / later than at present ... leading to a crushing of any remaining traffic. 

In summary

It's not the individual "north" and "south" that makes up the best use of the route and the needs of users; it's the flow of real people, in combination of where the want to go North and South from Melksham and Trowbridge IN ADDITION to the through traffic.   In some ways, the shorter the route the easier to organise campaigning over the smaller geographic area - but the through traffic there would be on the trains we want to run anyway. No easy answers here.

In spite of all the ideas I've spun, I am not a professional railway engineer, nor a planner.  But I know enough to be able to read, analyse, suggest - and I hope to come up with and talk about ideas that merit investigation in detail.    Thus  I'm sort-of concluding (for this post but NOT for ever) that a service from Salisbury to Swindon would get the highest use levels / the highest return or the least subsidy.   Other Southern points to consider would be Westbury, or BEYOND Salisbury (Southampton).  Other Northern points - Oxford or Chippenham.  My personal view is to remain very open through 9 (3 x 3) options that offers and to listen to further inputs on operational practicality, traffic level forecasts, etc.