Time changes, but HL1 can still follow
The origin of the idea for BZS was created back in 2007 - 2008 as I recall. I and Robo_killa joined back then as new and eager developers.The mod was called Bleach: Soul Slayers (BSS)
At that time the team was quite small and the talk about the engine started already before we got a coder.
At that time the Unreal engine was very popular from the Unreal 2004. One of the things which I had looked into scripting with. We knew that Unreal 3 was coming, but the way Unreal 3 should be modified would change from what we knew and since none had experience with Unreal development, it was quickly out of the question.
We had experience with the Half-life engine and HL2 was also a possibility. But at that time the community was large in the development of HL1 and HL2 was just getting started. Many coders would also say that HL1 was easier than HL2. Today HL2 has so much support that new coders can find the information they need compared to back then.
So HL1 was chosen for its simplicity and getting started on familiar grounds.
In 2010 BSS lost its coder and the code we had was not the final, so we started fresh with HL2 which we couldn't find any coders for. After a half year we found one which we didn't get long with.
We did also look at the crytek engine, but we didn't like the requirement of the buildings to be modelled as the crytek engine's building blocks lacks the functionality we had for mapping in HL.
In 2011 we chose to revert back to the HL1 engine and start over and even rethink the concept. Bleach: Zanpakutou Sensei (BZS) was born.
At this time Tacef was leader, coder and upcomming dad. I had enrolled Computer Science at the university and things got back on track.
When Tacef had to leave, I took over and continued on the code we had done with Roei.
So now we had a lot of content for HL1, both models (some from BSS) and code, so we chose to stay with HL1 despite various engines on the marked.
The gold source engine (HL1) is perhaps not the most ideal choice, at least HL2 would give us more possibilities for content.
However HL1 can still exceed its limits on some features, like models which reach outside the world box, and we have great opportunities for including third-party code(like physics and graphic effects) to Half-life without compatibility problems.
The way HL1 uses data files and code is very educational, as there is no sandbox editor as in Crytek and neither a fully development-environment as the Unity engine. This requires us to design correctly and plan new features carefully with every developer from the start, as the content needs to be linked together manually and be compatible with each other.
This is a good place to start, as this experience will be close to the one in the industry and the benefit is that our development tools aren't limited to one suite but can use a large variety of both commercial and free tools.
Another beneficial for the user, is that HL1 itself doesn't require a big gamer setup to run. It also gives our developer the opportunity to work with the mod and play the mod, on a computer that don't have the latest hardware.
BZS is a mod that don't have to be the best mod in the world, its goal is also with great focus on the opportunity to learn and adapt, which is one of the reason that the mod doesn't develop as fast as some others do.
We are all happy with HL1 as the selected engine for many of the above reasons, but unfortunately it also has the cost of not being a big part of the mainstream community; this however gives us the time to actually complete the development, instead of releasing a semi completed mod and retrieve a lot of critic from the community, like DayZ.