[Read description] Abandoned Project

This is a project I started over a year ago, and recently decided to abandon. I've decided to put it on the workshop for educational or inspirational reasons, since new mappers don't usually get to see failed projects.

The Project:

• The original idea behind this map was to condense the 'escape' section of the campaign into a single map.
• Destroying the neurotoxin tank was the first task, and that in its entirety made it into the map.
• There is also a turret factory section that I started building but didn't finish (that is in the map, but there is no gameplay).
• Each of these tasks would have been optional apart from the neurotoxin, and the map would culminate with an encounter with GLaDOS, where depending on how many of the tasks you completed, the fight would be easier or harder.

Reasons for abandonment:

• I've learned a lot since starting the project, and a lot of the early work is far below my current standard. In order to feel happy releasing this as a finished map, I would have to re-do most of it.
• The map is physically way too big. Hammer slows to a crawl, a full compile takes hours, and on top of that, the BTS doesn't justify it. What I've learned is best with BTS maps is to avoid the all too common 'void full of boxes' approach, and instead make the BTS area a large-ish closed room that hints at a larger world. The goal is to feel big without actually being big. This was the approach I took in my last chamber, 'A Peculiar Angle' and the approach that Valve takes in the campaign.
• I have other projects on the go at the moment, and I don't like having unfinished business.
• I tend to make maps according to obsessions that come in phases, and my current obsessions no longer line up with the ones I had when I began this project. In essence, I'm no longer as inspired by the idea as I was at the start.

What went well:

• I'm fairly happy with the neurotoxin tank, even though I had to put the pipe on backwards to get it to work with my map layout that I had already implemented. I think the idea of combining two separate tests (that can both be completed individually) in order to destroy the tank is also a nice idea that I may revisit.
• I like the opening chamber, where it is technically possible to miss the escape window, complete the very easy test, and move on to the next map in the queue without ever knowing what you missed.
• I spent a decent amount of time taking the neurotoxin tank from the campaign, tidying it up, and turning it into an instance I can reuse in future. It is a lot more complex than you would think (or than it should be, to be honest), and I'm glad that I can now easily use it in future when I want to.
• Although you can fall off the edge, I quite like the ratman den.

Obvious problems:

• Missing kill triggers for if you fall off the catwalk (so there are lots of ways to get stuck).
• Lack of clipping in many areas that need it.
• No significant optimisation.
• Generally poor level design.
• Lots of ugly examples of bad lighting or bad detailing.
• Many catwalks are floating and lead nowhere.

What I would do differently:

• Optimise as I go.
• Make use of cordons to speed up the workflow.
• Focus the scope to just one or two of the tasks, and not make either optional.
• I think a bossfight is probably the wrong conclusion to a map like this. What would be better is a play on the section in escape where GLaDOS tries to lure you with a fake exit, then hits you with neurotoxin if you fall for it. A version of this where the exit is real and GLaDOS's death traps have been disabled would be very funny, if rather out of character.
• The setting makes absolutely no sense. The weird pit with the stone and rebar textures don't mesh with the BTS floating chambers at all, and the reservoirs look silly at best.