I'm pretty sure DLL's are a windows only thing lol.
I am not sure because here is an example somebody is giving for a Linux machine.
I was just going to do bash calls for an example.
That would work also, its not like we need ultimate performance.
Could you send me some documentation on what exactly is going on here like what commands are being used?
1 Like
Weeeeeeeeeee
http://www.monodevelop.com
Tool I'm going to use. Easy to make QT and GTK stuff, then just add code shit. Optionally there is also gnome-builder but thats not in windows.
https://wiki.gnome.org/Apps/Builder
Then, I have problems getting prime to work for AMD cards. Theres also this finicky piece of shit called bumblebee for nvidia switchables and optimus. Optimally, have an app that just opens with sudo, displays your GPU info at the bottom that it can pull from lspci -v, maybe like a gpu-z style window, then have the features listed out in another tab.
https://wiki.archlinux.org/index.php/hybrid_graphics
https://wiki.archlinux.org/index.php/NVIDIA_Optimus
https://wiki.archlinux.org/index.php/PRIME
Shouldn't be hard to make, the problem will be if hardware gives issues. Example, my y40-70 doesn't actually allocate PCIe lanes for its dgpu so essentially its a macbook. I think theres other laptops from lenovo that are like this and hardware thats like an alienware might give issues as well. But, at the same time, useful tool that no one has made yet.
A bit late to the party, but I just thought I would throw an idea out there that I have been stewing over for quite some time:
Has anyone here seen Bountysource or something similar? I see lots of people venting about issues with certain things not working, fragmentation, too many distro's and so on, and I feel making a new distro is not helping in the long run. Starting at the bottom and fixing or making the things we want, from the ground up and open source, seems to be a method to consolidate and make the community better vs yet another fork/respin distro.
I really like the idea of crowd funding fixes and features to make Linux better vs donating to a large group and hoping they make what we want. We can have the power to create the future of open source software. I wrote a gigantic blog post about this which I need to edit and post when I get home this weekend, but I think it would be awesome if people here good at coding could check out Bountysource, or share any other places like it.
If we can create a positive atmosphere where people can work or donate towards one small goal at a time it can create a lot of momentum. People seeing their donations making real change, however small, will give incentive to donate again. Having larger bounties to attract talented people could certainly be beneficial. I'd like to hear what other people think of this.
2 Likes
I mean, to be honest, if we were going to help push a distro along I would say help ubuntu out with package management.
1 Like
Can we make a level 1 internet roaming bot?
2 Likes
Reporting in for the cause.
Unfortunately I won't be of much use on the development side of things, but I could contribute on the project management/business aspect. At the moment I have plenty of time to take ownership of a project when some centralization is needed and github (or something similar) won't suffice anymore.
As a non-professional enthusiast of IT I really have to agree with @KleerKut that currently the barrier-to entry for using linux is way to high, and I personally would love to see some consolidation within the world of linux.
On the one side this would entail making the use of linux more user friendly and intuitive in the form of bug fixes, GUI wrappers to replace the need of the terminal, etc.
On the other side good documentation (i.e. guides, etc) is needed so people know what they are doing (this is why I love the L1T videos, the pfsense ones are an excellent example of this).
In order to really make a difference in the long-term I think it is not only important to make solid contributions to the software (like the excellent suggestions in this topic), but also to keep things recognizable and centralized for the lesser informed/newbies. A L1T distro would be the pinnacle of this ideology, but this will require some serious commitment in order to make it a game changer.
A more realistic example could be a L1T repository which is well-maintained and accompanied with accessible documentation, bug reporting, etc. (after receiving the blessing by @wendell and his partners in crime off course).
To give an idea of what I am meaning: I recently started experimenting with unraid, and instead of researching/trying each docker image, I just look what "linuxserver" has available because I know these images are relatively stable and come with excellent documentation.
In order to really make a change and keep things organized, I think the following would help
- Sorting different use cases (office desktop, gaming desktop, home NAS, etc)
- Selecting the right distro for each of these use cases
- Selecting the right additional software packages where applicable.
- Testing these "final" combinations, inventory what steps need to be taken to make it a pleasant user experience
- maintain/alter/create the software and create/update the accompanying guides, how-tos etc.
After a slow start I really started diving into linux the last couple of months and although I love learning and researching new things (thanks L1T) it really gets frustrating sometimes spending hours and hours of reading obscure forums and guides just to get one thing working. I guess this is the topic I chose to vent these frustrations and secretly hope I could do something about it.
Anyway, I will be keeping an eye on this topic and am looking forward to contribute. Apologies if the above does not make any sense, its late over here.
2 Likes
EVERYONE STOP WE'RE USING LOLCODE
3 Likes
The thing I keep running into is that there are not any well designed, general, and well implemented complex data structures in C/C++. Graphs? They've all been trash. Sparse Matrixes? None are performant. Matrix multiplication? Prepare thy anus. Oh, you wanted an average, sum, or information for a whisker plot? Not in C you're not! General case alignment? Let us skip a few steps and give you the shared substring because you wouldn't have used that intermediate data anyways, right? Skip lists? The STL would be embarrassed. Suffix array? Oh, you wanted test suites.
Much of my time has just been doing not half assed (see STL quality) like data structures and then just the high level algorithms which aren't hard when you have the structure to rely on. From a research perspective, this is a big roadblock and most researchers don't have the time or skill to do it right, and certainly not the money to hire out to do it.
I'd like some help with these.
This was one of the things i read about the C "standard" library was that its kinda crap when you need something efficient and works properly.
So, I've been toying with talking about my "current" project for a while now, but I guess this is the perfect time to do this.
I've been prototyping a backup program, similar to duplicity or CrashPlan, but the goal is to be better. It's still in the proof of concept stage, but I'd like to get ideas, feedback, discussion, whatever from people.
This project spawned out of me being frustrated as hell with the way that CrashPlan makes it difficult for you to connect to headless (freenas) daemons to do configuration and monitoring. I figured there had to be a better way to implement everything anyways, so I thought I'd try to build something better.
Proof of concept is written in Python 3, but I'm open to changing language to pretty much anything that's not Java or Perl. Python's just easy to prototype in. If someone's interested in working on this with me, definitely let me know.
I'm also not attached to the name. I just needed something to name the repo, so if anyone has a better idea for names, please let me know. I'm not a creative person.
code is here:
I know my code is shite, I'm not really a good programmer.
So here's the premise:
Daemon runs in the background and every so often will scan for changes and store the changes in the storage engine. The storage engine is more of a director than anything else. The backup engine passes the data chunk off to the storage engine where it decides where to store the data. Based on your preferences, the goal is to be able to store data locally (external hard drive, mounted network drive, etc...), on a storage daemon available over the network or in a cluster. The cluster is something that I thought would be cool, it would allow you to keep multiple copies of your data on multiple machines and scrub and repair if something happens to the backup. Let's say you have a local NAS and a VPS with 4TB space on it somewhere. Deploy the cluster daemon on both the NAS and the VPS, configure them and set your laptop to target either of them. From there, they default to storing 3 copies (if possible) of the data. Your laptop will back up directly to the machine you're targeting, then the target will handle the clustering. From there, when you periodically scrub the backups, if you have data that's not in 3 copies somewhere or if you have data that's not passing checksums, you can pull the data from another system and rebuild your data.
Since the system stores changed chunks, it's also capable of versioning data. So you can go back in time in case you've deleted something or made an oopsy.
Sorry for the massive wall of text, I set out to talk briefly about this and wound up writing more than I intended.
The C libc stuff is problematic, but the STL is only in C++, and tends to be one of the better implemented libraries around. Now, that's not a high bar, but they are standard for a reason. I think you're thinking back to the glibc implementation of sin() or something, which is awful, and also a thing that should be improved. I'm just more interested in the higher level data structures.
Since we are in need of a wiki, I propose that we knock out two birds with one stone.
We have a community project to develop the wiki software that will shall use here at L1T.
Edit: I'm gonna do this when I get home. Gonna make a RoR project. Will update with the github link.
2 Likes
Here is another proposal. The forum is open source and has a number of reported bugs (here) and things people would want. That's something the community could probably work on as well.
3 Likes
A like assist counter would be cool.
In video games, if you help someone, you get an assist. The same should be for the forums. So if someone replies to your comment and they get likes, you get your assist ticker ( which would be located under your page ) increased.
I like this idea
I second.
Contributing to the forum is likely the best place to put in energy/time.
@Eden I think you meant to link to the open issues?
I was on my phone and to lazy.
but we have bugs reported here https://forum.level1techs.com/c/l1techs/Bug-report
and discourse forums here https://meta.discourse.org/c/bug
code is here https://github.com/discourse/discourse
its mainly ruby and javascript
1 Like