I wouldn't be frowning! Obviously it would be nice to have the intention of looking at all dates, no matter how long it takes us to actually get there
, but if there are reasons why you know that's never going to happen then stating that at the outset seems sensible. Perhaps you could put together a single page for the website listing the resources that you are aware of for the later period so that people know what survives and where they can find it? That in itself obviously wouldn't make a one place study but would be useful as a supplement to the main work or as a placeholder for future work.
You may also get someone who comes along who is also interested in Melrose and is keen to work on periods and records that aren't your priority? Some of the OPS sites I have listed have a local history group behind them rather than just one person, which always sounded pretty luxurious to me! Perhaps in your aims you could indicate whether you are interested in working collaboratively, each focusing on their own timeframe?
I think you've raised an interesting point about the census though. The accessibility of the census has changed so much over the last five years that simply transcribing it isn't where we add value, it's in the accuracy of those transcriptions and, depending on how you make your data available, emphasizing the context of an entry amongst its neighbours or linking it to other pieces of information - a different experience than just searching for a name in a census database and having it spit out the results. Focusing on other less-accessible records from the non-Victorian era, especially if that's a more interesting period to you, is probably contributing more to the field of family history at this point in time.