So there's been a lot said in this thread and I won't bother quoting most of it, but I will go over what we do know.
First, this LoS "exploit" as it were, has two separate variants, one due to uneven land, and one due to foundation types. Both have a lot of evidence suggesting that it was possible to do during pre-T2A up to present day (as of a few months ago at least).
The first type, which is commonly known here as the "tower bug" allows players to target any LoS spell under a tower foundation, hitting any player within the house at virtually any location by simply targeting the desired location, standing a few tiles back, and using either the Razor or in-game last target function (in-game LastTarget function was introduced with the second T2A client in August 1999). This particular behavior is a manifestation of the same type of bug outlined in a January 1998 newsgroup post about an old castle break in exploit involving teleporting into the courtyard by standing back a few tiles, and targeting one of the visible tiles in the courtyard, which was fixed by a November 1997 patch preventing players from teleporting within 3 tiles of any multiobject (this code is still in place on OSI servers as of a few months ago). This behavior can also be replicated in the demo, and can also be done on live servers today, which shows a consistency of mechanical behavior stretching back to, at minimum, 2 months after the game came out.
While we know exactly how this particular "exploit" works, the issue with attempting to fix this behavior is that we don't know what kind of unintended consequences that this behavior will have on other normal functions in game. There's also the very relevant fact that this particular behavior is still possible now, and so far as we know, has been possible since the beginning of the game. Granted, the use of this behavior to kill players standing inside houses was rendered irrelevant in 2006 with a patch that prevented players from damaging anything with LoS spells that they couldn't see, so it may not have been relevant for them to attempt to fix it. However, the fact that this particular use of this behavior was possible for over 7 years when I figured out what caused it in less than a week is telling.
The second type, which is known for its use on houses near uneven landscape, is a manifestation of the exact same code that allows players to teleport onto the roof of places like the Britain bank, or onto the roof of any flat structure. This behavior, which was implemented by Derrick from code decompiled from the demo core by Batlin, was only possible once that code was implemented here. Tests via the UO:98 project (the demo client does not contain the necessary functions to produce the desired results, but late T2A clients could do it just fine) and on live OSI servers show that identical behavior can be achieved in both environments to those currently found on UOSA.
This is an even more significant issue than the last one, as this deals with the core LoS function. While we could certainly choose another LoS function, we would be messing with the very core of how countless functions in the game determine the behavior of various functions from whether someone can shoot an arrow at a given angle, to whether a player can cast a spell on someone on severely uneven ground. It is also virtually impossible for us to come up with an equation for LoS that will have the same behavior in all other circumstances while preventing players from utilizing the same technique for teleporting onto the roof of a flat building, but only in instances when attempting to target a house from an uneven location. There is a virtual guarantee that any LoS code we would implement would break other things in favor of fixing this, which wouldn't serve to fix anything but simply shift the problem elsewhere.
There's more that can be said regarding whether this is or isn't an exploit (see: teleporting on roof vs throwing potions. Which is an exploit, and how can you differentiate them if one is and the other isn't?), or whether we have any evidence that OSI actually banned anyone for doing this, but I've stated my position on this elsewhere, and searching the forums for my posts on this behavior will doubtless bring you to them.
As a final note on this subject, this behavior isn't going to be changed, unless you can convince Derrick to have a change of heart on the subject, and then get him to actually implement something else.
Now for the part that I will quote:
Mens Rea wrote:With respect, Kaivan didn't play T2A UO, how would he know what it was all about anyway.
I actually did play during T2A. I started in September 1999, and played through until AoS was introduced.