Although it could be argued that, if the Up exit from Rescue Esmeralda to Ballroom East is a valid route, then so is using WRITETYPER to access any room! Perhaps that is even more valid, as Matthew obviously intentionally implemented WRITETYPER, whereas the Esmaralda/Ballroom shortcut may have been an unintentional mistake.
I would agree that an argument could be made for the validity of a game being completed using WRITETYPER. After all, it's something that is there, so one may argue there is no reason why it shouldn't be used. I agree with this.
On JSW Central, I mention the highest documented scores in MM games / best completion times in JSW games (where pertinent). The condition is that they should be achieved without cheating (by which I mean using the WRITETYPER or POKing / modifying the game) and (in case of JSW games) without unnecesary loss of life (if an item can be collected without loss of life, it *must* be collected without loss of life; kamikazing some items could yield a significant time gain).
I could envisage (but it's hypothetical at the moment, because I have no time for this, JSW Central needs other kind of development first) having four categories for best completion times in JSW:
- No unnecessary loss of life, no use of WRITETYPER;
- No unnecessary loss of life, but the use of WRITETYPER permitted;
- Lose as many lives as you want, but no use of WRITETYPER;
- Lose as many lives as you want plus the use of WRITETYPER permitted.
The last category should produce the best completion times.
However, I would NEVER even consider telling the players they should not perform some manoeuvres which are possible to do in an unmodified game file. As explained above, for me it's a kind of designer-versus-player game. If the designer made some mistake and left some unintentional loopholes, it is perfectly OK for the player to exploit them, because they are there.
Although being a little bit of a designer myself, I have a strictly player-centred perspective on these matters. The player loads a game and plays it. He doesn't have to know - and in many, many cases doesn't know - *anything* about the designer, who he/she was, what his/her intentions were, nothing. They may not even know the designer's name. All they get is a product, a game, which they face and need to solve. Whatever can be done in an unmodified game is perfectly valid.
Now, how does WRITETYPER fit in here? It is something that can be done in an unmodified game, so yes, I would agree that in a sense it is perfectly valid.
However, it introduces a new quality into the way the game is played, one that can affect the completion time significantly. It is for that reason that best completion times achieved using WRITETYPER (where possible) and not using it cannot be classified together - because we are talking about two different ways of playing the game.
It's a bit like if the designer places a cheat menu up front in the game (which does happen in some hacked versions of the original JSW) and provides, e.g. invulnerability to guardians. It is an option which the designer has provided in an unmodified product (from the point of view of the player, let's say I'm taking about an original game, not a hacked version), so in that sense it's perfectly valid to use. However, *obviously* playing the same game with invulnerability to guardians and without it are two very different modes of playing the game. Therefore, if the best completion times were to be discussed, there would *have to* be two separate categories for these two modes of playing the game. It's exactly the same with WRITETYPER.
Please note that an additional thing is that cheats such as WRITETYPER are perfectly definable in whether they have been used or not. However, if one started speculating about the author's intentions, you would get into a grey area of uncertainty. Did the author really intend it this way or not? And even if he did, and failed, as mentioned before - I believe it's his problem, not the player's, because it's not about the author's intentions, but about what can - or cannot - be done in the finished product, which is the game the player loads into his emulator / Vega / real hardware / GBA / whatever.
One more point: it's interesting that MM has a built-in cheat which allows the player to explore the game / practise playing each room easily, but the use of which prevents the appearance of the swordfish sign at the game's completion. In a sense it's a shame that Matthew Smith didn't implement something similar in JSW (which would be really easy to do with a simple check, I think) which would be: if WRITETYPER has been used, Maria doesn't disappear even if all items have been collected (or she does disappear, but the toilet run doesn't start upon reaching the bed - no "true" ending).