ERAM (Open Source RAM Disk) Development

[Zero3K]

It would be nice if someone could get it to work on Windows 10 (64-bit). Its driver and source code can be downloaded by going to https://github.com/Zero3K/ERAM.

[f0dder]

Did you write this, or get the source from elsewhere and dump it on GitHub?

Are there anything particularly interesting about this RamDisk, compared to ImDisk?

[Zero3K]

1. I got the source from the latest archive of it that is available on its homepage (which is located at http://hp.vector.co.jp/authors/VA000363/).
2. Its faster according to the benchmark I put on the repo and smaller in size.

[f0dder]

So you just snatched that code and stuck a GPLv3 on it?

Size doesn't really matter for a driver (as long as it's not grotesque), speed does - I don't see a benchmark on the GitHub repo, though, only a single CrystalDiskMark screenshot, which doesn't tell anything without a point of reference.

[Zero3K]

1. Yes.
2. Ok. You can find a comparison of some RAM Disks at http://phase.s214.xrea.com/module/techreport/ramdisk/.

[f0dder]

You can't just take other people's work and stick an arbitrary license on it.

The domain from your link is blocked by uBlock, so...

[Zero3K]

I have attached the benchmarks as a PDF.

[wraith808]

I have attached the benchmarks as a PDF.

More than the benchmarks, the problem is the licensing.  As f0dder said:

You can't just take other people's work and stick an arbitrary license on it.

It's not even whoever sources that page, as shown in the disclaimer: http://hp.vector.co.jp/authors/VA000363/tech/eramv68.htm

Text below:

Views on ERAM modified version of 2ch

Last updated 2008.7.13
68 Mr. at 2ch (?) Set aside also serves as a survival prove views on ERAM for WindowsNT / 2000 / XP / 2003 source difference is published.
Seems to be there was one that is copyright Donokono of problems seen but, (as long as does not contain copyrighted material of others) you can safely source publication of all or a difference with respect to ERAM.
On for people should be readily available execution module, it can be safely be published the source bundled with or separately from binary difference. At your own risk If you are using the binary difference.
It seems people are stay that is published a binary module, but it is a little. I think, but it is not intended to force the deletion.
If possible (68 Mr. version to the body) might also be where I am allowed to ask the source coalesce, but there is no spare capacity (in 2008.7.13 time). Operation check environment of over4GB is no at hand.

[Zero3K]

So then why offer the source code for modification?

[wraith808]

So then why offer the source code for modification?

Setting a license is not the same as offering source code- especially something like GPLv3.  That particular license, for better or worse, is a virus.

You can't choose to set that on someone else's source. It says something specific about what you're trying to do, and you can't decide that for someone else.



It's not originally your code.  It's not originally your project.  You can contact the developer to clarify their position.  But not being able to find that developer, and/or not getting a response does not leave it open for you to decide that.  See Circle Dock for the example of what bad things can happen when you try to do so.  This is a big deal for a lot of people.  And when things are a big deal, making unilateral decisions for others isn't a good thing to do, no matter what you might think.  Their reactions might surprise you.

[Zero3K]

I removed the License file.

[f0dder]

I removed the License file.

If you want to go through with this, you should probably contact the original author(s) and get some clarification on what license they want.

When something is released without an explicit license, it falls under normal copyright, and can't be considered OpenSource - so it's a bit dangerous contributing to, and doesn't belong on GitHub. I know this might seem pedantic, but keep the CircleDock fiasco in mind...

And you should add an AUTHORS file that make it very clear who and where the code originally came from.

[Zero3K]

1. I think he's either dead or gave up on it a long time ago (its been 10+ years since the latest version was made).
2. Please link me to a pge that explains this fiasco.
3. I will do that soon.

[f0dder]

1. You should still try contacting the author if you haven't done so already - he might not care, but it's really not a good idea getting into copyright trouble.
2. It happened here on DoCo, and it wasn't pretty.

Also, the benchmark PDF you attached seems to be slightly old, and if ImDisk is present, I missed it. Given the somewhat murky status of ERAM, you're probably better off contributing to ImDisk if your interest is OpenSource. If you just want a good, small, fast and  free (beer) RamDisk, get SoftPerfect.

[Zero3K]

1. I sent him an email regarding the source code being on Github.
2. Oh, ok.
3. I still think that there's more speed to get out of this driver when it works properly on a 64-bit OS.

[wraith808]

http://www.donationcoder.com/forum/index.php?topic=21947.msg372109#msg372109

http://www.donationcoder.com/forum/index.php?topic=23480.msg214378#msg214378

http://www.donationcoder.com/forum/index.php?topic=23705.msg215157#msg215157

It was ugly, and it turned DCers into something I hadn't seen on DC. So that's the reason we're being a bit pedantic about it.  I personally don't want to see that happen here again.

[Zero3K]

Yeah, that was bad. So, I take it that no one wants to try improving it?

[wraith808]

Yeah, that was bad. So, I take it that no one wants to try improving it?

Improving... what?

[Zero3K]

Yeah, that was bad. So, I take it that no one wants to try improving it?

Improving... what?

The driver. It would be nice if the BSOD I saw on a 64-bit OS was fixed so that way I can see if the speed is improved over what I got when I benchmarked it in a 32-bit OS.

[f0dder]

3. I still think that there's more speed to get out of this driver when it works properly on a 64-bit OS.

You seem somewhat obsessed about this particular RamDisk - why? Even according to benchmark you attached, it doesn't seem to be anything special... and it has a nasty CPU spike (might be a glitch - the benchmark doesn't strike me as being particularly thorough).

For people to want to spend their time improving something, there has to be some motivation. The uncertain license status is a turn-off, and in general it just seems more worthwhile to contribute to something not dead, like ImDisk.

[Zero3K]

3. I still think that there's more speed to get out of this driver when it works properly on a 64-bit OS.

You seem somewhat obsessed about this particular RamDisk - why? Even according to benchmark you attached, it doesn't seem to be anything special... and it has a nasty CPU spike (might be a glitch - the benchmark doesn't strike me as being particularly thorough).

For people to want to spend their time improving something, there has to be some motivation. The uncertain license status is a turn-off, and in general it just seems more worthwhile to contribute to something not dead, like ImDisk.

1. The benchmark is out of date. It might be faster now thanks to improvements in technology.
2. Its been a few months since its last update though.

[skwire]

2. Its been a few months since its last update though.

Unless I'm mistaken, I'd daresay a few months is much better than the years since this ERAM code has been updated.

[Zero3K]

I agree. But still, its source is out there now for anyone to develop it further. Btw, anyone want to post their own benchmarks of it in order to see if its faster on a real machine?

[Zero3K]

Well, the supposed impossible happened. The developer of it responded to me. Here is the email he sent me:

遅くなりすみません。
・私は英語は読めますが、満足には書けません。
・オリジナル版の著作権は放棄していません。
・over4GB版(f*ck版)は私の作成ではないので、ソースは手元にありません。
・公開しているソースコードを元に、改造するのは自由です。
・GPLではないので、ソースコード公開を強制してはいけません。

--- http://www.excite.co.jp/world/english/
It becomes late, excuse me.
* I can read English, I can't write English satisfactorily.
* A copyright of an original edition isn't given up.
* over4GB edition (f*ck edition) isn't my making, so a source isn't in hand.
* It's free to remodel the source code which is being exhibited into origin.
* It isn't GPL, so you aren't supposed to force source code openness.
---

[Zero3K]

I managed to fix the BSODs thanks to the help of a member of community.osr.com. Now I am getting some after removing the 4 GB limit and testing one that has a size of 6 GB. You can see the pull request with the changes I did by going to https://github.com/Zero3K/ERAM/pull/10.