Well, I really hope you can see this through... sounds very time consuming.
As you suggest, I've spent a great deal of time on it so far. I have written and rewritten the "stories" involving the Ventaria. I have developed a quest outline and project plan. Then I started designing weapons and quickly grew frustrated. I have Adobe Masterworks Suite CS4 (which includes Photoshop), as well as several professional plugin packages (Topaz and Nik's software packages to name two), and I've have a couple of Photoshop classes, so I'm not a complete newb to image or texture editing. Still, I couldn't make swords as beautiful as the ones StarX made. (I can retextures a sword; the trick is making it look, well, perfect).
As I've said, finding StarX's swords was fortuitous. That saved me a lot of time.
Anyway... I would refrain from sending the player back and forth through the same dungeon. I personally wouldn't like to delve into the same dungeon twice unless there's large portions of the dungeon still to explore behind locked doors. I would also try to make the creatures unique and poweful instead of the standard spawns or a mix of both. A few smaller boss mobs here and there can really spice up any dungeon crawl. Respawns, that's another story. I prefer the feeling that I actually achieved something when I cleared out that dungeon and sent those undead back into their graves. If they just keep respawning it doesn't really feel heroic or of any impact at all. If it would be possible I would even change the way the dungeon "feels" into something less dark and threatening after it has been cleared, i.e. by changing the light, the fog or the ambient sounds. The common mine and elven ruin xyz might have respawns but quest dungeons which are of a certain importance should feel unique.
About the deep dungeons I liked the way Bethesda approached this. Usually in the larger dungeons you can unlock secret doors which provide quick connection between the different levels, but those secret doors can only be opened from the other side. That way you have to explore the dungeon at least once to open the secret passages but at the same time you can use the quick route the next time you enter.
Yes, I've never felt good about sending players back and forth. However, in one case it's going to be absolutely necessary. In other cases, it makes the story richer, and gives the Ventaria depth. I suppose the quests could be given to the player as a request (Will you do this for me after I'm dead ...) and then the player can choose do them or not.
What you're suggesting with creatures would certainly increase the creation time for the dungeons. Currently, the dungeons rely on spawn points to create enemies. That's easy. Just put a spawn point in where you want a creature. This allows me to use Oblivion spawns points that mods like FCOM and MMM rely upon, so players with those mods will get the all the creatures they add.
But that is the catch: the dungeons are like every other dungeon the player enters. And the creatures respawn. To change this, I created the catacomb design, with has several rather unique features - in terms of lighting, etc. Using clouds and lights settings, I can create a different "feel" for the dungeons.
As I already noted, the Crypts/Tombs follow the same layout. This is a limitation based upon the Bethesda tileset I chose. It can really only be constructed in one way, with some very minor changes. To make the the ten tombs look and feel different, I've put different things in them, played some tricks with the lighting, and so on.
Someone else also suggested using gates or secret doors to allow access to other parts of the dungeon to explore, thus making it easier for the player to return to the dungeon, if necessary. It would be like running up a flight of stairs, sort of.
Tthe heart of what you're suggesting - the removal of spawn points in favor of specifically placed creatures - would definitely make the dungeons more unique. Creatures could be tailored to the personality of the Ventaria, to reinforce the theme of their story. And that would certainly take much more time to develop as I would want something than just standard Oblivion creatures and I would have obtain permissions to use some of MMMs stuff, and so on. Another issue would be making the mod flexible enough so that the player wouldn't feel as if he had to be right at Level X to enjoy the mod. So from a design perspective, it might be nicer but more limiting. From a development perspective, it would take a lot longer. Definitely worth thinking about, though.
For the impatient players who just want the swords... add the IDs to the readme. I wouldn't even comment on such requests. If people don't want the dungeon crawl but only the swords... just let them spawn the weapons via the console.
But twenty hours of gameplay probably requires more than a year to develop for one person. Honestly. Without any artificial time consuming loops you let the player jump through (which tends to aggravate players) to prolong the playtime you can be happy if you achieve a playtime of one to two hours. And that's even more than some commercial DLCs can claim which are created by whole teams.
What I would suggest is that you build your mod in stages or episodes. One dungeon (and sword) at a time, each with its own story and background but at the same time the later episodes build upon and require the previous ones.
That statement about hours is probably true. I don't fast travel. I can't fast travel. I use a mod that adds some cool map features but allows me to disable fast travel. I do use Arthmoor's excellent Cyrodill Travel Services, but my characters still do a LOT of walking. Also, I disable the green triangle and tend to forget others still use it - and feel quite passionate about using it.
By example: I usually complete the Mages Guid Recommendation quests in about two to three in-game weeks. I travel. I rest. I get sidetracked helping the citizens of Cyrodiil.
I suppose if someone relied upon fast travel, they hit all the dungeons, fight their way through, get all the swords ... I don't know. I can never judge time or difficulty. I remember when I started playing FCOM. I had just learned how to sneak and use poisons. I fought my way through Ft. Blueblood at Level 6 - then someone told me it shouldn't be attempted until Level 12. So sure, the right player will probably just sail through this.
You say stages and I think of Dumas and dikeens, writing their novels one chapter at a time. Still ... there might be something too that. I wouldn't feel rushed and I could add one or two more side quests that would make the entire mod story stronger. Each release could have it's own title ... (yes, I think it terms of story.

) with the final release bringing the entire story together. I will seriously consider that. In fact, I will probably follow some for that design.
Edit: By the way... how would you like the idea to create a new Katana/Sword mesh with unique new textures? As much as I like StarX' weapons they are already freely available. A mod of that scope should award the player with something completely new.
Edited: The discussion of this topic will take place off forum. StarX is on vacation and I owe him the courtesy of making no decision without his input.
~ Dani ~