Big changes for the Nexus Mod Manager and the introduction of Tannin42, our new head of NMM development

HadToRegister · October 14, 2016

In response to post #43231020. #43231160, #43234190, #43235645 are all replies on the same post.

Reveal hidden contents

Quote
prinyo wrote:

On 10/14/2016 at 7:27 AM, Tannin42 said:

In response to post #43228135. #43228440, #43228710, #43228800, #43229750 are all replies on the same post.

Reveal hidden contents

Quote
Mebantiza wrote: So will it be safe to say, this 'new' MM have will none of the features of NMM, and all the features of MO. IoW, just a reskinned and renamed MO? As bad as NMM 6x has been, there are things about NMM that make people want to use it, well, the pre- 6x versions anyhow. How is forcing MO and its structure on everyone, even those that don't really want it, a benefit?

I use NMM 56.1, and do NOT require virtual installs. Its a feature, for some, but not a REQUIRED one, and I dont want it forced on me. 56.1 is old, crashes constantly, and is not all that stable tbh. But its still far preferable to anything NMM 6x, and have no need to have MO forced upon me. Not going to argue which one is 'superior', that is pointless. NMM is straight-forward enough to do what it has to, and did it well enough, well at least untill they tried to steal some of MO thunder. Then it went to complete and utter $yht. Up until that point, it wasn't perfect, but it did the job. NMM 6x added a needless feature that no one asked for, and never worked properly. Had NMM retained its pre-VM structure and improved that, things would likely be fine with it even now.

So now, the 'solution' it seems, is make everyone use MO whether they like it or not...

Quote
HadToRegister wrote: Having used both, I like MO, and I also like NMM, but only pre 0.62 NMM, as 0.62 and above just acted like malware and destroyed most of our mod lists

NOWHERE in this thread has it been said that the new NMM will have NONE of the features of NMM and/or ALL of the features of MO.

NOWHERE has the subject of "Forcing MO on everyone" even been discussed.

If you use NMM 56.1 (For whatever hipster anti-something reason), then keep using it, but stop complaining about something that you're not only NOT going to use, but are also so misinformed about, that you were unable to make even one factual statement about the plans for NMM/MO for the future.

Quote
xyon71 wrote: Why go right to negativity??
I think it is very UNsafe to say it is just going to be a "re-branded MO"
It appears to me that they are wanting to produce a simple, yet powerful NEW tool for everyone to use with varying degrees of complexity based on need.
I didn't read anywhere that they are "forcing" anything, I read that they are taking all of the feedback from BOTH tools to incorporate the best features into 1 "best" tool.
I switched from NMM to MO 2 years ago, and found it simply wonderful to be able to try out different mods in different orders without breaking my base game.
I will admit there was a learning curve, but not too steep, and I am far from an advanced modder.
I for 1 am excited to see what these talented minds can come up with.
For whatever this or that, I say all of them are in it because they want to give people the best modding experience possible, so lets give them a chance before we jump on our band-wagons and shoot harpoons because we "think" we already know what they are going to do...

Quote
archerarcher wrote: Same here. In times where disk space is cheaper than fuel I think this virtual install feature is completely obsolete. I don't like it and I don't like the problems it makes with certain mod installations.
I use NMM 56.1 for Skyrim too, it's very stable to me, no problems at all, really. Okay, I track all my (700~) mods and over the years my documentation has grown into a complex website but I know exactly what to overwrite and what not and I know where my files are and I don't care about some gigabytes disk space more or less for my favorite game that I am playing/ modding since Nov 2011.

So what should a future mod mager be like?

1. Manage mod files
That's the absolute priority. Everey feature that affects this should be removed or be avoided.

2. Simple to use, simple to manage
No need for virtual installs, no installation of mods directly by clicking on a button on a website. I need control! Everyone needs control. Is there anybody out there who installs mods via website button? I inspect EVERY file before I use a mod so you should too if you don't want to loose track of your installment.

That's what I want from a mod manager and that's why I use NMM 56.1 and not MO

Quote
moriador wrote: @archerarcher,

Control. Precisely.

The more lines of code that come between me and the stuff I want to use, the more difficult it becomes to fix what ought to be minor problems. Instead of simply locating the problem file or reference and fixing it, you have to contend with the installation software and all of its quirks. Instead of replacing a single bad mesh, for instance, you may end up having to uninstall an entire mod, create a new mod archive with the new mesh, and install it all over again.

The more control you take away from users, the more those users have to rely on mod authors (and mod manager programmers) to fix every single little problem.

But there seems to be a trend in software to create bigger and bigger walls of code between the user and the actual assets/files that user wants to use. I see it in 3D render software, image editing software, anything that relies heavily on user created content really.

Which is completely backwards, IMO. The more imperfect the assets, the more DIRECT control the user needs to have in order to fix those imperfections. And the more an industry relies on user generated content, the more imperfections there will be.

The idea of creating code that allows a user to click a single magic button is great indeed! But only if that code depends on assets that have been through a very rigorous quality assessment process which requires very strict standardization. If those assets are potentially riddled with bugs, as any user created content will be, then you can't wall the user off from direct control or their only option when things go wrong will be to scream at the developers.

I see a lot of screaming at mod authors and mod manager programmers going on in Nexus comments (and the forums of many other industries). But I guess that's the price you pay when you try to make things *too* easy for the user.
It's no rebranded MO, it's a fresh start. "Controversial" MO features like the virtual filesystem may exist as options / extensions but they will not be defaults or requirements.

With MO I was happy to develop a tool that would only appeal to a small crowd. It was always intended to be complement the existing solutions like OBMM/NMM or wrye bash, not to replace them.

Now we're writing a modding tool for all Nexus users so obviously the approach will be different. I'm not ignorant of the problems with MO but with MO I had one target audience, now it's a different one.
We try to make the new mod manager attractive to advanced & MO users through extensibility, not by doing the same again and hoping the majority of users will suddenly like it better.

This sounded really good until I read the comments on the STEP forum where there is this: "The default setting will probably work similar to what NMM is doing currently, using symlinks or hardlinks into the game directory because this is simply more robust "

This is bad news. This is exactly what nobody wants. This is exactly what people want to get rid of.
When people talk about "the way NMM works" they mean pre-0.6. If the new manager uses the current NMM logic then everybody loses.
It was obvious that the choices you make about this will upset one of the two crowds. But upsetting both of them at the same time... hmm

Quote
Tannin42 wrote: Ok, can you explain what's so bad about the *link approach?

Quote
Gruftlord wrote: If I had to make a guess, I'm sorry to tell you, that I think it looks like quite a few of MO's users don't actually understand how the other mod managers work and how they do the things they glamour MO for. If anything, it's a testament to MO's more accessible documentation, direct advertisement of advanced features or user interface differences.

Quote
Exoclyps wrote: MO is simple. I can just open up the folder for my mod and mess with it without a second thought. I can even manually add mods by adding another folder.

Some link approach that seems to upset a lot of people does not sound promising to me. It sounds complicated and I get the impression I won't have the same simple accessibility I do today.

I'll still have MO on my harddrive so I'm not really losing anything atm, but yeah, I won't convert to any other virtual system when there is one already that is really clean and easy to use.

Quote

prinyo

This is bad news. This is exactly what nobody wants. This is exactly what people want to get rid of.

Whoa, hold on there, I don't remember electing you to represent everybody, and you certainly don't represent me.

remlnx · October 14, 2016

Congratz Tannin, i wish you can be free to use your awesome skills and NMM will be totally rebuild from scratch, and not just work to improve the discrete ugly semi-useless manager it is now.

Sad to see the best manager (MO) being dropped after so many years it made my life easier, and clean my games installations.

JimboUK · October 14, 2016

In response to post #43210430. #43215310, #43215490, #43215920, #43216655, #43217565, #43217935, #43218210, #43222660, #43225585, #43227190, #43227900, #43227920 are all replies on the same post.

Reveal hidden contents

Quote
jim_uk wrote: Please can we get a mode for us old stick in the muds who still do everything manually and only want something to enable mods and change the load order? I'm still using 0.52.3.

Quote
kingtobbe wrote: This works perfectly for that. It doesn't disable mods beyond unchecking the esp though. Of course if it did, it wouldn't be simple anymore :)

http://www.nexusmods.com/fallout4/mods/13671/?

Quote
vram1974 wrote: You're never too old to save time. If a veritable moron like myself can figure out mod managers then anybody can.

Quote
jim_uk wrote: It's nothing to do with figuring them out, it's that I don't need all the other stuff, I've been installing mods since 2002, I know what I'm doing and prefer to do things myself.

@kingtobbe I used something similar for FO4, it saved me from upgrading and risking breaking my older games.

Quote
Arthmoor wrote: I'm with jim_uk on this one. If I were to be convinced to use this new NMM, I'd want it to be straightforward and without all the file system virtualization stuff. And it would need to be at least as robust about handling the Data folder as Wrye Bash is now.

Oh, and please, for the love of Talos, leave the BSA unpacking to other tools where it belongs!

Quote
HadToRegister wrote: So if you're still using 0.52.3 then just continue using it then, problem solved, as you're not even using NMM for anything except to enable/disable mods and change the load order.
You're not even using it to it's fullest potential, so you'll be able to use that version forever.

Quote
jim_uk wrote: It won't work for newer titles, this months Skyrim Shiny Edition being one of them.

Quote
PirateZ86 wrote: @jim_uk: Please don't listen or reply to morons who have no common sense.

Quote
HadToRegister wrote:

Quote

PirateZ86 9 kudos 557 posts
@jim_uk: Please don't listen or reply to morons who have no common sense.

Wow, so you go straight to the ad hominems because someone has a different opinion than you?

Jesus christ dude, chill out.

Quote
ShalabiRogue wrote: @arthmoor This could be a dumb (or too long to answer) question but what do you have against file virtualization? For me the benefits are massive; quick and easy profiling, clean skyrim folder, quick and easy file priority, and MO lets me use Wrye Bash for bash patches and such even between profiles. I can't even think of a negative point.

Without MO I couldn't have both Enderal and Skyrim installed right now and quickly choose which one I want to play.

Quote
Arthmoor wrote: My main objection is having an intermediate layer between my OS and my game that is under the control of neither one and prone to bugs because neither one has been subjected to the appropriate levels of QA.

Plus I just don't see the point considering Wrye Bash has not had issues with "dirty data folders" since the BAIN installers module was perfected. Proven technology that gives you the kind of control MO users want without injecting code into the task to accomplish it.

IMO if you want something sufficient to separate Enderal from Skyrim you should be prepared to sacrifice the extra install space for it. Drive storage is cheap, and nothing beats a hard separation when it comes to guaranteeing one can't contaminate the other.

Also, I don't need to jump through hoops for each tool I want to use in order to get the virtual file system to notice it. Which ironically includes not having to faff about with that to make Wrye Bash work to do bashed patches with :P

Quote
moriador wrote: Yes, exactly.

There doesn't seem to be much point in using a file virtualization system if the CK can't see it.

Quote
ColdHarmonics wrote: While I can sympathize with this position, I think anybody who has taken the 5 minutes it takes to install and learn MO will soon understand its profound advantages and near-complete lack of disadvantages, 100% of the time. As someone who very much likes doing things myself, MO is a tool that lets me do things myself better. But, to each their own.

Read what people have written, it's nothing to do with learning how to use it, don't assume people who prefer to do things differently are somehow ignorant.

NCRForever · October 14, 2016

In response to post #43213170. #43213650, #43214685, #43224605, #43231310, #43231435, #43232450, #43232890, #43235755 are all replies on the same post.

Reveal hidden contents

Quote
stebo104 wrote: Mod Organizer is still a million times better then NMM.

Quote
renthal311 wrote: NMM is so made that without a tutorial, even a novice can install selected mods, I'm Modder, and I do complicated things, but I could not install even one mod MO no tutorial, a huge difference, NMM is very user-friendly, simple, fast, is the best and will continue :)

Quote
Thallassa wrote: To install a mod in MO is exactly the same as in NMM. Double click the download file, activate the mod. Play.

MO just has ton more power.

Quote
matortheeternal wrote: renthal: Installing a mod with Mod Organizer is no more difficult than with NMM. This talk about it being difficult to use is all just a bunch of mythology.

It's literally exactly the same. Click download with NMM link on Nexus Mods, double click download in the downloads list/tab, check mod in mod list. That's it. :|

EDIT: Didn't refresh page, Thalla beat me to it.

Quote
renthal311 wrote: I do not deny, my first encounter with MO was bad, I just had this feeling that I can not anything to install, now already to me unnecessary, does not install mods, only their own, to test the game, so enough for me NMM, otherwise I am a man, who reluctantly changed, something that is already good for me, that is, NMM :) :)

Quote
renthal311 wrote: Interestingly, my first encounter with NMM were that: 1. I downloaded a mod,, 2 installed, I had no idea about it, but as a '' green '' worked the first time, the application suggests itself, the design of NMM is perfect even for beginners and New urzytkowników, Modding communities, Nexus :)

Quote
Exoclyps wrote: Thallassa and matortheeternal, I think the problem might be the NXM thingy, gotta make sure it links to MO. And installing SKSE might be a bit more effort.

But if the MM would be able to use the NXM thingy out of the box (which it ofcourse will) and detect various things like SKSE, TES5Edit and the likes automatically I think it'd streamline things further.

I personally use MO and takes me just mere seconds to setup, no problem. But there is a few things to keep in mind when you set it up.

Quote
archerarcher wrote: @ stebo104:
Your aren't seroius, are you?

Quote
Exoclyps wrote: archerarcher, but it is better than NMM. Waay better.

Installing SKSE with MO is as easy as KISSing(Keep It Simple Stupid). Similar to installing it with NMM. You drag and drop the loose files into the install folder, then using 7zip or WinRAR, right click the Data folder, name it to SKSE scripts(If you want) and 7zip/zip/rar it. Load MO and add it like a downloaded mod. Click the box to install and you're done.

See? Easy as KISSing.

shazara · October 14, 2016

Thanks for the update and you seem to be making all the right moves in terms of moving forward with the right team and approach. Since you are making development more formal now - it would likely also be a good idea to also have a more formal method for Feature Requests vs reading a bajillion posts in a bajillion threads all across the nexusmods site. Any word on what that the plans might be on that front?

HadToRegister · October 14, 2016

In response to post #43232045. #43232355, #43233385, #43233940, #43234660, #43235000, #43235590 are all replies on the same post.

Reveal hidden contents

Quote
prinyo wrote:

On 10/14/2016 at 9:03 AM, Tannin42 said:

In response to post #43231020.

Reveal hidden contents

Quote
prinyo wrote:

In response to post #43228135. #43228440, #43228710, #43228800, #43229750 are all replies on the same post.

Reveal hidden contents

Quote
HadToRegister wrote: Having used both, I like MO, and I also like NMM, but only pre 0.62 NMM, as 0.62 and above just acted like malware and destroyed most of our mod lists

NOWHERE in this thread has it been said that the new NMM will have NONE of the features of NMM and/or ALL of the features of MO.

NOWHERE has the subject of "Forcing MO on everyone" even been discussed.

If you use NMM 56.1 (For whatever hipster anti-something reason), then keep using it, but stop complaining about something that you're not only NOT going to use, but are also so misinformed about, that you were unable to make even one factual statement about the plans for NMM/MO for the future.Mebantiza wrote: So will it be safe to say, this 'new' MM have will none of the features of NMM, and all the features of MO. IoW, just a reskinned and renamed MO? As bad as NMM 6x has been, there are things about NMM that make people want to use it, well, the pre- 6x versions anyhow. How is forcing MO and its structure on everyone, even those that don't really want it, a benefit?

I use NMM 56.1, and do NOT require virtual installs. Its a feature, for some, but not a REQUIRED one, and I dont want it forced on me. 56.1 is old, crashes constantly, and is not all that stable tbh. But its still far preferable to anything NMM 6x, and have no need to have MO forced upon me. Not going to argue which one is 'superior', that is pointless. NMM is straight-forward enough to do what it has to, and did it well enough, well at least untill they tried to steal some of MO thunder. Then it went to complete and utter $yht. Up until that point, it wasn't perfect, but it did the job. NMM 6x added a needless feature that no one asked for, and never worked properly. Had NMM retained its pre-VM structure and improved that, things would likely be fine with it even now.

So now, the 'solution' it seems, is make everyone use MO whether they like it or not...

Quote
xyon71 wrote: Why go right to negativity??
I think it is very UNsafe to say it is just going to be a "re-branded MO"
It appears to me that they are wanting to produce a simple, yet powerful NEW tool for everyone to use with varying degrees of complexity based on need.
I didn't read anywhere that they are "forcing" anything, I read that they are taking all of the feedback from BOTH tools to incorporate the best features into 1 "best" tool.
I switched from NMM to MO 2 years ago, and found it simply wonderful to be able to try out different mods in different orders without breaking my base game.
I will admit there was a learning curve, but not too steep, and I am far from an advanced modder.
I for 1 am excited to see what these talented minds can come up with.
For whatever this or that, I say all of them are in it because they want to give people the best modding experience possible, so lets give them a chance before we jump on our band-wagons and shoot harpoons because we "think" we already know what they are going to do...

Quote
archerarcher wrote: Same here. In times where disk space is cheaper than fuel I think this virtual install feature is completely obsolete. I don't like it and I don't like the problems it makes with certain mod installations.
I use NMM 56.1 for Skyrim too, it's very stable to me, no problems at all, really. Okay, I track all my (700~) mods and over the years my documentation has grown into a complex website but I know exactly what to overwrite and what not and I know where my files are and I don't care about some gigabytes disk space more or less for my favorite game that I am playing/ modding since Nov 2011.

So what should a future mod mager be like?

1. Manage mod files
That's the absolute priority. Everey feature that affects this should be removed or be avoided.

2. Simple to use, simple to manage
No need for virtual installs, no installation of mods directly by clicking on a button on a website. I need control! Everyone needs control. Is there anybody out there who installs mods via website button? I inspect EVERY file before I use a mod so you should too if you don't want to loose track of your installment.

That's what I want from a mod manager and that's why I use NMM 56.1 and not MO

Quote
moriador wrote: @archerarcher,

Control. Precisely.

The more lines of code that come between me and the stuff I want to use, the more difficult it becomes to fix what ought to be minor problems. Instead of simply locating the problem file or reference and fixing it, you have to contend with the installation software and all of its quirks. Instead of replacing a single bad mesh, for instance, you may end up having to uninstall an entire mod, create a new mod archive with the new mesh, and install it all over again.

The more control you take away from users, the more those users have to rely on mod authors (and mod manager programmers) to fix every single little problem.

But there seems to be a trend in software to create bigger and bigger walls of code between the user and the actual assets/files that user wants to use. I see it in 3D render software, image editing software, anything that relies heavily on user created content really.

Which is completely backwards, IMO. The more imperfect the assets, the more DIRECT control the user needs to have in order to fix those imperfections. And the more an industry relies on user generated content, the more imperfections there will be.

The idea of creating code that allows a user to click a single magic button is great indeed! But only if that code depends on assets that have been through a very rigorous quality assessment process which requires very strict standardization. If those assets are potentially riddled with bugs, as any user created content will be, then you can't wall the user off from direct control or their only option when things go wrong will be to scream at the developers.

I see a lot of screaming at mod authors and mod manager programmers going on in Nexus comments (and the forums of many other industries). But I guess that's the price you pay when you try to make things *too* easy for the user.
It's no rebranded MO, it's a fresh start. "Controversial" MO features like the virtual filesystem may exist as options / extensions but they will not be defaults or requirements.

With MO I was happy to develop a tool that would only appeal to a small crowd. It was always intended to be complement the existing solutions like OBMM/NMM or wrye bash, not to replace them.

Now we're writing a modding tool for all Nexus users so obviously the approach will be different. I'm not ignorant of the problems with MO but with MO I had one target audience, now it's a different one.
We try to make the new mod manager attractive to advanced & MO users through extensibility, not by doing the same again and hoping the majority of users will suddenly like it better.

This sounded really good until I read the comments on the STEP forum where there is this: "The default setting will probably work similar to what NMM is doing currently, using symlinks or hardlinks into the game directory because this is simply more robust "

This is bad news. This is exactly what nobody wants. This is exactly what people want to get rid of.
When people talk about "the way NMM works" they mean pre-0.6. If the new manager uses the current NMM logic then everybody loses.
It was obvious that the choices you make about this will upset one of the two crowds. But upsetting both of them at the same time... hmm
Ok, can you explain what's so bad about the *link approach?

Even if we assume that the problems that now exist with half-install and half-removing of mods leaving stray files in the data directory are caused by old and un-optimized code, there are still enough reasons that turn such a system into a nightmare for the advanced users and are now forcing them to use 0.56.1 as the last version of NMM that is actually usable. And those reasons have nothing to do with the quality of the software itself but the problems the concept of linked virtual install creates. Having the files in one place is logical and easy to work with (like MO and the old NMM), having ghosts of files in more places makes the experience quite limiting, confusing and annoying.

1. Backups - it is not possible to make an usable backup independently of the manager. What do you copy - the files in the data folder or the files in the VirtualInstall? Or both? Whatever you do there is no meaningful way to use such a backup. You are forced to make a copy of the two as this is the only way to use it and that means - 2 times longer do make/apply the backup, 2 times more space for the backup, 2 directories to backup instead of one and most importantly - twice the size of the game folder after the backup is used. This is a huge deal as most players put those games on small SSD drives. If you have 50GB of mods and want to use a backup then your game will now consume a total of 100GB. And even if the drive space is not a big problem, having several games inflated like that is not really acceptable.
Because yes - when you install a mod and create links for all the files then you are not using twice the space - you have one and same file with two different addresses. So 1k file in data = 1k file in VI = 1k total. But when you copy the files and put them somewhere else the bond will break and you will end up with two independent "real" files. And yes - you need to make and use a backup of both data and VI folders. If you only use the data folder the game will work but the mod manager will become useless. If you only use a backup of the VI then there will be no mods in the game... So you are forced to endure the double backup and inflated drive space usage.

2. Freedom, predictability and flexibility (and way less chaos) - it is surprisingly easy to break the link and end up with two different files. This limits your possibility to reuse self-modified mod assets on a new install, create confusion when you edit a file that you don't immediately realize is not tied to it's counterpart anymore, forcing you to constantly think about which files are tied and which not and which copy you are editing. The "schizophrenic" state an install like this exists in is extremely confusing and adds way more complications than it is worth.

3. Learning curve. It is been mentioned many time on the forums that the learning curve for NMM 0.6* is steeper than MO. And it is true. Introducing this to the "basic" variation of the software seems strange.

4. For what? People would probably accept all this complications and problems if the would see the benefit such a system brings. However there is none. The predictability, usability and stability of the "one file" install logic is sacrificed for what... For profiles? Let's be honest, everybody who wanted profiles was already using MO. Now if you want to create the basic modding tool and with options to add modules and upgrade, are there Profiles in that very basic tool? There is no "greater good" here. It is complication for the sake of the complication. None of the potential benefits of the virtualization are realized by the linked virtual install, only the negatives.

I would suggest starting a new thread about this and letting all people that still use the old NMM versions tell their reasons. Because I'm pretty sure none of them would be using this new manager if uses the same logic that made them stay away from the new versions anyway. I understand from the point of view of the MO crowd the phrase "advanced NMM users" is an oxymoron, but in reality there are advanced users who have experience in both pre- and after- 0.6 NMM.

Quote
Exoclyps wrote: IMO, what I think would be best is to give two options. One where you can use a basic version like old NMM and one where you can create MO like profiles, using the system that is used there.

Quote
Tannin42 wrote: Thanks for the detailed answer. half-installations / half-removals are of course a problem of the implementation not of the concept. There is no reason to assume the same problems would appear in a fresh implementation

1) I can't speak too much about NMM 0.6x but the way I intend to implement it you would only ever have to backup the mod directories managed by the mod manager and you will then be able to restore your mods on a different system or after reinstalling the game.

2) I agree. If you run, say TESVEdit on a file that belongs to a mod, afterwards you have a real file in the data directory with no way of telling if it once belonged to a mod. But I don't see how this is different from a regular file install? Unless you have a "manifest" of which file belongs to which mod you can't keep track and this works the same for links and real files.
Even MO has problems with that which resulted in the "overwrite" directory.
I have two possible alternative approaches for virtual installs in my head that might solve this problem but I'd rather not go into too much detail getting hopes up before having verified it can be implemented.

3) Again, this is an implementation detail that I don't intend to repeat. With the new manager the intended behaviour is that things just work and you are only confronted with profiles when you actually start using them.

4) I disagree, the virtual install stuff makes a lot of things easier especially when dealing with file conflicts: If you do a regular file install, mods overwrite each other. When you now remove / disable a mod that, upon installation, overwrote a file from another one, you have to restore that file. A mod manager has to keep track of which files exist in which mod in a "manifest" (the InstallLog in NMM) and it has to keep the archives around so it can restore the files.
This is massively error prone because if this manifest gets damaged you're screwed, the mod manager completely loses its knowledge of previous installs.
MO had far fewer problems with stuff like this - not because it used vfs magic but because it kept mods in separate directories from the very beginning. It never tried to keep a "manifest" valid and up-to-date between versions - it simply didn't need one.
"Virtual Installs" make the mod management way more robust no matter how they are implemented.

I agree that this is something that needs to be discussed in detail. There are multiple points I don't want to just decide without having heard opinions.
But my impression is that your gripe with the VirtualInstalls stuff in NMM 0.6x is more with how it was implemented than the underlying idea.

Quote
Gruftlord wrote: I really like how you are handling the comments and criticism coming from both sides. Kudos to you! From what i read i get a strong feeling that your vision for the new manager is indeed one, that satisfies the needs of either side.

Quote
Contrathetix wrote: Just to provide another user opinion: having the possibility to use virtual mod installation like with Mod Organizer (as in, no symbolic links or anything else even semi-permanent created in the game directory) would be great, and I am sure many people would value such an option. Not only does it make actual installation a lot faster (not everyone has SSDs), but it also removes a lot of hassle and uncertainty when it comes to "what do I have in my game folder, exactly?" There is a certain element of... ease attached to virtual installations. A peace of mind. The way MO works is exactly what pleases my inner control freak (of sorts) when it comes to mods. Having each mod in its own folder is also very useful for mod makers, when everything related to one project can be kept in its own folder, with no need whatsoever to hunt for files that get buried somewhere in the depths of the data folder. The virtual installation feature is also useful for saving disk space (not everyone has terabytes of it laying around). It is also the one single reason I switched from Wrye Bash to Mod Organizer for installing mods.

Now I have never ever used NMM, but my experiences with the MO virtual install have been brilliant, no complaints whatsoever. If something should be the future of mod installation, that something should be virtual installation! It is fast, it is easy to understand (on a general level), it keeps the data folder clean, it looks very difficult to break (if not impossible, never tried it), and the user does not really need to pay much attention to its existence (excluding having to remember to launch tools through MO).

If it is somehow possible to offer a completely virtual install option like with MO, then that would be handy. Tannin, you still have USVFS, and a project of that complexity must have taken some time, thinking and grey hairs to actually come to be, so instead of leaving it to gather dust, continuing its use with the new manager might be one way to carry it forward. Maybe if the mod installation + process launch parts of the new manager could be bundled into an add-on like block that could be changed?

Come what may, I am still looking forward to the new mod manager. :)

Edit: Talking about the USVFS-powered virtual installation, not anything that requires writing symbolic links or any other stuff somewhere (because, in my opinion, that sort of defeats the idea of a virtual installation). USVFS should be the future of mod installation.

Edit 2: Maybe it really could be handy to make sure people actually know what "virtual installation" in MO means. Reading the posts, it seems that people might think it uses symbolic links or other link stuff... which is a bit odd. If NMM uses symbolic or hard links and calls it "virtual", then maybe that is confusing people to a certain extent? Maybe? If people actually understood how the MO/USVFS virtual mod installation works, then maybe they would not be so much against it.

Quote
turulo wrote: 1) all fine until:
a) there is a bug in NMO (nexus mod organizer) and all that strategy goes to the floor because you cannot backup/move the game to another place, as you don't have idea what the hell is inside the game directory.
b) you need to reinstall the NMO or use another version of NMO (like a personal compiled version because you cannot wait to get some bugs fixed) and the other NMO doesn't have the internal database of mods installed for that game.
c) use a program that is not programmed to handle the virtual file system such as bodyslide and breaks all the NMO assumptions that the file in the virtualinstall is the original.

The virtual install should be optional, because bugs happen and people that don't want to deal with that extra layer of complexity should be allowed to exist.

Another option is to have the virtual install and an option to convert a virtual install to a hard install (I mean, something to convert those links to the real file).

Quote
Exoclyps wrote: turulo, I got the feeling you never used MO, but rather got annoyed with the recent NMM? Since there is no way a bug or anything could mess things up in MO since it uses simple folders to keep track of things.

Quote

turulo 13 kudos 234 posts
1) all fine until:
a) there is a bug in NMO (nexus mod organizer) and all that strategy goes to the floor because you cannot backup/move the game to another place, as you don't have idea what the hell is inside the game directory.
b) you need to reinstall the NMO or use another version of NMO (like a personal compiled version because you cannot wait to get some bugs fixed) and the other NMO doesn't have the internal database of mods installed for that game.
c) use a program that is not programmed to handle the virtual file system such as bodyslide and breaks all the NMO assumptions that the file in the virtualinstall is the original.

The virtual install should be optional, because bugs happen and people that don't want to deal with that extra layer of complexity should be allowed to exist.

Another option is to have the virtual install and an option to convert a virtual install to a hard install (I mean, something to convert those links to the real file).

I agree with making a virtual install OPTIONAL and I like the option to convert a virtual install to a hard install, but the first part of your posts definitely sounds like you've never used MO at all, because if you did, you'd know that each mod is put in their own folder so the files never physically overwrite each other and other mods, which means you can painlessly rearrange your load order without having to keep track of what files you overwrote and have to replace/restore etc.
And each mod folder was clearly named, not like the current version of NMM that names the folders by their mod ID

HadToRegister · October 14, 2016

In response to post #43224730. #43229715, #43231195, #43231225, #43234520 are all replies on the same post.

Reveal hidden contents

I like what someone previously suggested but without the 'x'

NeMO = Nexus Mod Organizer Edited October 14, 2016 by HadToRegister

Ethreon · October 14, 2016

In response to post #43222435. #43222490, #43225970 are all replies on the same post.

Reveal hidden contents

Yea, you don't play more than a game, nobody plays them. Maybe you need to re-read the article and see what the numbers are.

Thallassa · October 14, 2016

In response to post #43224730. #43229715, #43231195, #43231225, #43234520, #43237900 are all replies on the same post.

Reveal hidden contents

I think NeMO is the best one I've heard, but I'm still partial to "NOMM" (not quite sure how that acronym works out but it's cute).

Thallassa · October 14, 2016

In response to post #43232045. #43232355, #43233385, #43233940, #43234660, #43235000, #43235590, #43237830 are all replies on the same post.

Reveal hidden contents

Quote
prinyo wrote:

On 10/14/2016 at 9:03 AM, Tannin42 said:

In response to post #43231020.

Reveal hidden contents

Quote
prinyo wrote:

In response to post #43228135. #43228440, #43228710, #43228800, #43229750 are all replies on the same post.

Reveal hidden contents

Quote
HadToRegister wrote: Having used both, I like MO, and I also like NMM, but only pre 0.62 NMM, as 0.62 and above just acted like malware and destroyed most of our mod lists

NOWHERE in this thread has it been said that the new NMM will have NONE of the features of NMM and/or ALL of the features of MO.

NOWHERE has the subject of "Forcing MO on everyone" even been discussed.

If you use NMM 56.1 (For whatever hipster anti-something reason), then keep using it, but stop complaining about something that you're not only NOT going to use, but are also so misinformed about, that you were unable to make even one factual statement about the plans for NMM/MO for the future.Mebantiza wrote: So will it be safe to say, this 'new' MM have will none of the features of NMM, and all the features of MO. IoW, just a reskinned and renamed MO? As bad as NMM 6x has been, there are things about NMM that make people want to use it, well, the pre- 6x versions anyhow. How is forcing MO and its structure on everyone, even those that don't really want it, a benefit?

I use NMM 56.1, and do NOT require virtual installs. Its a feature, for some, but not a REQUIRED one, and I dont want it forced on me. 56.1 is old, crashes constantly, and is not all that stable tbh. But its still far preferable to anything NMM 6x, and have no need to have MO forced upon me. Not going to argue which one is 'superior', that is pointless. NMM is straight-forward enough to do what it has to, and did it well enough, well at least untill they tried to steal some of MO thunder. Then it went to complete and utter $yht. Up until that point, it wasn't perfect, but it did the job. NMM 6x added a needless feature that no one asked for, and never worked properly. Had NMM retained its pre-VM structure and improved that, things would likely be fine with it even now.

So now, the 'solution' it seems, is make everyone use MO whether they like it or not...

Quote
xyon71 wrote: Why go right to negativity??
I think it is very UNsafe to say it is just going to be a "re-branded MO"
It appears to me that they are wanting to produce a simple, yet powerful NEW tool for everyone to use with varying degrees of complexity based on need.
I didn't read anywhere that they are "forcing" anything, I read that they are taking all of the feedback from BOTH tools to incorporate the best features into 1 "best" tool.
I switched from NMM to MO 2 years ago, and found it simply wonderful to be able to try out different mods in different orders without breaking my base game.
I will admit there was a learning curve, but not too steep, and I am far from an advanced modder.
I for 1 am excited to see what these talented minds can come up with.
For whatever this or that, I say all of them are in it because they want to give people the best modding experience possible, so lets give them a chance before we jump on our band-wagons and shoot harpoons because we "think" we already know what they are going to do...

Quote
archerarcher wrote: Same here. In times where disk space is cheaper than fuel I think this virtual install feature is completely obsolete. I don't like it and I don't like the problems it makes with certain mod installations.
I use NMM 56.1 for Skyrim too, it's very stable to me, no problems at all, really. Okay, I track all my (700~) mods and over the years my documentation has grown into a complex website but I know exactly what to overwrite and what not and I know where my files are and I don't care about some gigabytes disk space more or less for my favorite game that I am playing/ modding since Nov 2011.

So what should a future mod mager be like?

1. Manage mod files
That's the absolute priority. Everey feature that affects this should be removed or be avoided.

2. Simple to use, simple to manage
No need for virtual installs, no installation of mods directly by clicking on a button on a website. I need control! Everyone needs control. Is there anybody out there who installs mods via website button? I inspect EVERY file before I use a mod so you should too if you don't want to loose track of your installment.

That's what I want from a mod manager and that's why I use NMM 56.1 and not MO

Quote
moriador wrote: @archerarcher,

Control. Precisely.

The more lines of code that come between me and the stuff I want to use, the more difficult it becomes to fix what ought to be minor problems. Instead of simply locating the problem file or reference and fixing it, you have to contend with the installation software and all of its quirks. Instead of replacing a single bad mesh, for instance, you may end up having to uninstall an entire mod, create a new mod archive with the new mesh, and install it all over again.

The more control you take away from users, the more those users have to rely on mod authors (and mod manager programmers) to fix every single little problem.

But there seems to be a trend in software to create bigger and bigger walls of code between the user and the actual assets/files that user wants to use. I see it in 3D render software, image editing software, anything that relies heavily on user created content really.

Which is completely backwards, IMO. The more imperfect the assets, the more DIRECT control the user needs to have in order to fix those imperfections. And the more an industry relies on user generated content, the more imperfections there will be.

The idea of creating code that allows a user to click a single magic button is great indeed! But only if that code depends on assets that have been through a very rigorous quality assessment process which requires very strict standardization. If those assets are potentially riddled with bugs, as any user created content will be, then you can't wall the user off from direct control or their only option when things go wrong will be to scream at the developers.

I see a lot of screaming at mod authors and mod manager programmers going on in Nexus comments (and the forums of many other industries). But I guess that's the price you pay when you try to make things *too* easy for the user.
It's no rebranded MO, it's a fresh start. "Controversial" MO features like the virtual filesystem may exist as options / extensions but they will not be defaults or requirements.

With MO I was happy to develop a tool that would only appeal to a small crowd. It was always intended to be complement the existing solutions like OBMM/NMM or wrye bash, not to replace them.

Now we're writing a modding tool for all Nexus users so obviously the approach will be different. I'm not ignorant of the problems with MO but with MO I had one target audience, now it's a different one.
We try to make the new mod manager attractive to advanced & MO users through extensibility, not by doing the same again and hoping the majority of users will suddenly like it better.

This sounded really good until I read the comments on the STEP forum where there is this: "The default setting will probably work similar to what NMM is doing currently, using symlinks or hardlinks into the game directory because this is simply more robust "

This is bad news. This is exactly what nobody wants. This is exactly what people want to get rid of.
When people talk about "the way NMM works" they mean pre-0.6. If the new manager uses the current NMM logic then everybody loses.
It was obvious that the choices you make about this will upset one of the two crowds. But upsetting both of them at the same time... hmm
Ok, can you explain what's so bad about the *link approach?

Even if we assume that the problems that now exist with half-install and half-removing of mods leaving stray files in the data directory are caused by old and un-optimized code, there are still enough reasons that turn such a system into a nightmare for the advanced users and are now forcing them to use 0.56.1 as the last version of NMM that is actually usable. And those reasons have nothing to do with the quality of the software itself but the problems the concept of linked virtual install creates. Having the files in one place is logical and easy to work with (like MO and the old NMM), having ghosts of files in more places makes the experience quite limiting, confusing and annoying.

1. Backups - it is not possible to make an usable backup independently of the manager. What do you copy - the files in the data folder or the files in the VirtualInstall? Or both? Whatever you do there is no meaningful way to use such a backup. You are forced to make a copy of the two as this is the only way to use it and that means - 2 times longer do make/apply the backup, 2 times more space for the backup, 2 directories to backup instead of one and most importantly - twice the size of the game folder after the backup is used. This is a huge deal as most players put those games on small SSD drives. If you have 50GB of mods and want to use a backup then your game will now consume a total of 100GB. And even if the drive space is not a big problem, having several games inflated like that is not really acceptable.
Because yes - when you install a mod and create links for all the files then you are not using twice the space - you have one and same file with two different addresses. So 1k file in data = 1k file in VI = 1k total. But when you copy the files and put them somewhere else the bond will break and you will end up with two independent "real" files. And yes - you need to make and use a backup of both data and VI folders. If you only use the data folder the game will work but the mod manager will become useless. If you only use a backup of the VI then there will be no mods in the game... So you are forced to endure the double backup and inflated drive space usage.

2. Freedom, predictability and flexibility (and way less chaos) - it is surprisingly easy to break the link and end up with two different files. This limits your possibility to reuse self-modified mod assets on a new install, create confusion when you edit a file that you don't immediately realize is not tied to it's counterpart anymore, forcing you to constantly think about which files are tied and which not and which copy you are editing. The "schizophrenic" state an install like this exists in is extremely confusing and adds way more complications than it is worth.

3. Learning curve. It is been mentioned many time on the forums that the learning curve for NMM 0.6* is steeper than MO. And it is true. Introducing this to the "basic" variation of the software seems strange.

4. For what? People would probably accept all this complications and problems if the would see the benefit such a system brings. However there is none. The predictability, usability and stability of the "one file" install logic is sacrificed for what... For profiles? Let's be honest, everybody who wanted profiles was already using MO. Now if you want to create the basic modding tool and with options to add modules and upgrade, are there Profiles in that very basic tool? There is no "greater good" here. It is complication for the sake of the complication. None of the potential benefits of the virtualization are realized by the linked virtual install, only the negatives.

I would suggest starting a new thread about this and letting all people that still use the old NMM versions tell their reasons. Because I'm pretty sure none of them would be using this new manager if uses the same logic that made them stay away from the new versions anyway. I understand from the point of view of the MO crowd the phrase "advanced NMM users" is an oxymoron, but in reality there are advanced users who have experience in both pre- and after- 0.6 NMM.

Quote
Exoclyps wrote: IMO, what I think would be best is to give two options. One where you can use a basic version like old NMM and one where you can create MO like profiles, using the system that is used there.

Quote
Tannin42 wrote: Thanks for the detailed answer. half-installations / half-removals are of course a problem of the implementation not of the concept. There is no reason to assume the same problems would appear in a fresh implementation

1) I can't speak too much about NMM 0.6x but the way I intend to implement it you would only ever have to backup the mod directories managed by the mod manager and you will then be able to restore your mods on a different system or after reinstalling the game.

2) I agree. If you run, say TESVEdit on a file that belongs to a mod, afterwards you have a real file in the data directory with no way of telling if it once belonged to a mod. But I don't see how this is different from a regular file install? Unless you have a "manifest" of which file belongs to which mod you can't keep track and this works the same for links and real files.
Even MO has problems with that which resulted in the "overwrite" directory.
I have two possible alternative approaches for virtual installs in my head that might solve this problem but I'd rather not go into too much detail getting hopes up before having verified it can be implemented.

3) Again, this is an implementation detail that I don't intend to repeat. With the new manager the intended behaviour is that things just work and you are only confronted with profiles when you actually start using them.

4) I disagree, the virtual install stuff makes a lot of things easier especially when dealing with file conflicts: If you do a regular file install, mods overwrite each other. When you now remove / disable a mod that, upon installation, overwrote a file from another one, you have to restore that file. A mod manager has to keep track of which files exist in which mod in a "manifest" (the InstallLog in NMM) and it has to keep the archives around so it can restore the files.
This is massively error prone because if this manifest gets damaged you're screwed, the mod manager completely loses its knowledge of previous installs.
MO had far fewer problems with stuff like this - not because it used vfs magic but because it kept mods in separate directories from the very beginning. It never tried to keep a "manifest" valid and up-to-date between versions - it simply didn't need one.
"Virtual Installs" make the mod management way more robust no matter how they are implemented.

I agree that this is something that needs to be discussed in detail. There are multiple points I don't want to just decide without having heard opinions.
But my impression is that your gripe with the VirtualInstalls stuff in NMM 0.6x is more with how it was implemented than the underlying idea.

Quote
Gruftlord wrote: I really like how you are handling the comments and criticism coming from both sides. Kudos to you! From what i read i get a strong feeling that your vision for the new manager is indeed one, that satisfies the needs of either side.

Quote
Contrathetix wrote: Just to provide another user opinion: having the possibility to use virtual mod installation like with Mod Organizer (as in, no symbolic links or anything else even semi-permanent created in the game directory) would be great, and I am sure many people would value such an option. Not only does it make actual installation a lot faster (not everyone has SSDs), but it also removes a lot of hassle and uncertainty when it comes to "what do I have in my game folder, exactly?" There is a certain element of... ease attached to virtual installations. A peace of mind. The way MO works is exactly what pleases my inner control freak (of sorts) when it comes to mods. Having each mod in its own folder is also very useful for mod makers, when everything related to one project can be kept in its own folder, with no need whatsoever to hunt for files that get buried somewhere in the depths of the data folder. The virtual installation feature is also useful for saving disk space (not everyone has terabytes of it laying around). It is also the one single reason I switched from Wrye Bash to Mod Organizer for installing mods.

Now I have never ever used NMM, but my experiences with the MO virtual install have been brilliant, no complaints whatsoever. If something should be the future of mod installation, that something should be virtual installation! It is fast, it is easy to understand (on a general level), it keeps the data folder clean, it looks very difficult to break (if not impossible, never tried it), and the user does not really need to pay much attention to its existence (excluding having to remember to launch tools through MO).

If it is somehow possible to offer a completely virtual install option like with MO, then that would be handy. Tannin, you still have USVFS, and a project of that complexity must have taken some time, thinking and grey hairs to actually come to be, so instead of leaving it to gather dust, continuing its use with the new manager might be one way to carry it forward. Maybe if the mod installation + process launch parts of the new manager could be bundled into an add-on like block that could be changed?

Come what may, I am still looking forward to the new mod manager. :)

Edit: Talking about the USVFS-powered virtual installation, not anything that requires writing symbolic links or any other stuff somewhere (because, in my opinion, that sort of defeats the idea of a virtual installation). USVFS should be the future of mod installation.

Edit 2: Maybe it really could be handy to make sure people actually know what "virtual installation" in MO means. Reading the posts, it seems that people might think it uses symbolic links or other link stuff... which is a bit odd. If NMM uses symbolic or hard links and calls it "virtual", then maybe that is confusing people to a certain extent? Maybe? If people actually understood how the MO/USVFS virtual mod installation works, then maybe they would not be so much against it.

Quote
turulo wrote: 1) all fine until:
a) there is a bug in NMO (nexus mod organizer) and all that strategy goes to the floor because you cannot backup/move the game to another place, as you don't have idea what the hell is inside the game directory.
b) you need to reinstall the NMO or use another version of NMO (like a personal compiled version because you cannot wait to get some bugs fixed) and the other NMO doesn't have the internal database of mods installed for that game.
c) use a program that is not programmed to handle the virtual file system such as bodyslide and breaks all the NMO assumptions that the file in the virtualinstall is the original.

The virtual install should be optional, because bugs happen and people that don't want to deal with that extra layer of complexity should be allowed to exist.

Another option is to have the virtual install and an option to convert a virtual install to a hard install (I mean, something to convert those links to the real file).

Quote
Exoclyps wrote: turulo, I got the feeling you never used MO, but rather got annoyed with the recent NMM? Since there is no way a bug or anything could mess things up in MO since it uses simple folders to keep track of things.

Quote
HadToRegister wrote:

Quote

turulo 13 kudos 234 posts
1) all fine until:
a) there is a bug in NMO (nexus mod organizer) and all that strategy goes to the floor because you cannot backup/move the game to another place, as you don't have idea what the hell is inside the game directory.
b) you need to reinstall the NMO or use another version of NMO (like a personal compiled version because you cannot wait to get some bugs fixed) and the other NMO doesn't have the internal database of mods installed for that game.
c) use a program that is not programmed to handle the virtual file system such as bodyslide and breaks all the NMO assumptions that the file in the virtualinstall is the original.

The virtual install should be optional, because bugs happen and people that don't want to deal with that extra layer of complexity should be allowed to exist.

Another option is to have the virtual install and an option to convert a virtual install to a hard install (I mean, something to convert those links to the real file).

I agree with making a virtual install OPTIONAL and I like the option to convert a virtual install to a hard install, but the first part of your posts definitely sounds like you've never used MO at all, because if you did, you'd know that each mod is put in their own folder so the files never physically overwrite each other and other mods, which means you can painlessly rearrange your load order without having to keep track of what files you overwrote and have to replace/restore etc.
And each mod folder was clearly named, not like the current version of NMM that names the folders by their mod ID

Tannin: Regardless of what approach you take, there should be a mechanism to manage files created dynamically. If I run TES5edit and make a patch, or if Campfire saves a .json file with its settings, or I run FNIS for users.exe and make a new .hkx file... in all of those cases that file needs to be *managed*, not just dumped in a directory.

In MO it goes in the overwrite, which works well for me; I just manually put it into a new mod where I need it to go. However me doing that is what I can do as an advanced user. I think some kind of prompt that says "You just created this file! What would you like to do with it?" might work better for newer users.... not at all simple to code, but simple to use.

Regardless at no point should the system attempt to predict what I want to do with the file. It shouldn't go "Oh, there's a similar file in this folder, I bet it goes here!" MO does this sometimes which is frustrating (but not always, which is even more frustrating). Trying to read the user's mind only leads to unpredictable bugs; much better to force the user to figure out their own mind :P

Big changes for the Nexus Mod Manager and the introduction of Tannin42, our new head of NMM development

Recommended Posts

Link to comment

Share on other sites

Top Posters In This Topic

Popular Days

Top Posters In This Topic

Popular Days

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Recently Browsing 0 members