Talk:GNUnet

Screenshots

Any screenshots out there for GNUnet or its GTK interface?

GNUnet itself is a network, you can't get a screenshot of a computer network right. There are some trace tools to create a visual representation of the neigbours of your node. Maybe this could be used somehow.

The old GNUnet-gtk in 0.6 (currently used version) is very plain and isn't as end-user-friendly as it should be for an average Joe User. It is still unclear, if it will be ported to 0.7, completely rewriten. Also "3rd party" GUIs may apear.

You still have a good point. I'm just not sure, if we have a good answer. --Easyas12c 00:54, 1 May 2005 (UTC)

It would be possible to reuse the Network stack images (if there's no license on them contact us, it *could* be implicitly covered by AGPL3 and the GFDL but I'm not 100% sure right now), the LEGO representation of services, and newer screenshots of pre 0.11 release of gnunet-gtk. N. (Ngzero (talk) 13:32, 6 June 2018 (UTC))

That's a good idea: The current screenshot seems to be very outdated. A new screenshot of pre 0.11 release of gnunet-gtk certainly will be fresher and a better representation of the 'what you see is what you get'-experience of an average new user in the year 2018. So, making such a screenshot and adding it to the article would improve the article.

Another good idea would be to add 2 pictures to the article: 1st the aforementioned new screenshot placed anywhere within the article, and 2ndly a picture visualizing the routing paths of the GNUnet or anything like that displaying it as a network in the infobox - just like done in the wikipedia article for the Internet - as a replacement for the old screenshot. This 2 picture approach would do justice to the fact, that the GNUnet is not just an anonymous P2P application, but rather a replacement for the internet, an alternative network stack, but on the other hand also would give insight into the user experience, which is very important, as the GNUnet is not just an abstract theory, but practical - one can download and use it. — Preceding unsigned comment added by 178.201.144.69 (talk) 08:09, 10 June 2018 (UTC)

Long URL

The long example URL really stretches the page badly. Can we line-break it in any sensible way without breaking the example? Haakon 20:41, 15 January 2006 (UTC)

Most browsers can't swap line (automaticly) in the middle of a "word". Doing it by hand would break copy/paste. So I think we can do nothing, but bug browser developers and/or part take in free software browser development. --Easyas12c 23:21, 15 January 2006 (UTC)

Browsers aren't supposed to linebreak a long word in the middle -- that's not a missing capability, that's HTML conformance. We could perhaps insert a space in the middle of the URL to let it be linebroken. Yes, that would break copy/paste, but I doubt most readers will be interested in doing that. But I think the best thing to do is just truncate the URL in the middle with an ellipsis. People would still get an impression of how a URL looks. Haakon 11:53, 16 January 2006 (UTC)

Many simple text editors are able to do this. Why shouldn't browsers be? Does html specifications really say a user agent is not allowed to break a word to multiple lines, no matter how long it is? That would be a design flaw in html then. Then it is even more complicated. We'd have to complain to World Wide Web Consortium. --Easyas12c 19:14, 16 January 2006 (UTC)

I kind of have the same problem with our (GNUnet) documentation rewrite in Texinfo. I'm considering to simply mention the format length once (breaking it if necessary) and then abbreviating the URLs if they are not part of a functional example. This is also how long URLs and folders in the GNU Guix documentation are treated. Maybe that works for you too? N. (Ngzero (talk) 13:29, 6 June 2018 (UTC))

ecrs module urls

The text about ecrs module url in the article is very hard to understand, if not impossible. --Easyas12c 23:31, 15 January 2006 (UTC)

Project status

I feel the fact that gnunet is essentially unusuable is salient information if it is to be covered in wikipedia at all. I'm not sure how this information should be presented. At the very least perhaps a notation that the software is not "ready for results-oriented users at this time, although participating may help development" or some such.

Essentially I feel that as-is the gnunet.org website and this article by extension are misleading. I and others I know have installed gnunet software and had it working (transferring identifiable real traffic in response to queries and such). My experience (similar to that of about 4 other users) on 0.7.0b was that the software consumed huge amounts of memory, cpu, and network bandwidth for up to a week (in my case), but was not able to even successfully acquire the example document (GPL COPYING file) which is of course less than 20 kilobytes in size.

Thus, in practical terms, this software does not work yet. I believe this puts it in a different category from most entries of software in wikipedia. JoshuaRodman 07:16, 21 February 2006 (UTC)

It does work. It is just not yet very user friendly [1]. So you may have misconfigured it, misused it or simply had bad luck. It is true that GNUnet still has many unresolved bugs [2]. However this is implicitly stated in the version number. It is a defacto standard to call the first end user release 1.0.0. --Easyas12c 15:12, 21 February 2006 (UTC)

I believe that the software may work for some people in some cases. However, there were no 'misconfiguration' steps to make. The node was up and communicating. Valid data was being passed. I managed to provide a series of items to the network complete with metadata. What did not work was downloading anything at all. That is, some data was downloaded on the default GNU COPYING file, but around 4kb was downloaded over the first day of operation. It is possible that if if I was "luckier" it may have worked in a sane fashion. But the system was up and operational, but provided laughable functionality. I set it up primarily because other people asked me whether gnunet was a viable project, stating similar results, and I simply could not believe it would work so badly, assuming they had firewall problems and the like. But as it turns out gnunet was not able to achieve even basic functionality when correctly configured.

I certain use many software packages prior to their consideration as 'final'. ext2fs v0.4 springs to mind. I don't expect perfection, but I expect core functionality to be usable. JoshuaRodman 06:09, 18 March 2006 (UTC)

0.7.0c seems to work a lot better than 0.7.0b. --Easyas12c 11:46, 18 March 2006 (UTC)

0.7.2c works really good. cpu and disk utilization are fine now ( resource limiting of cpu and network work well). 14:35, 04-oct-2007 —Preceding unsigned comment added by 217.7.99.234 (talk) 12:36, 4 October 2007 (UTC)

17:20, 3 June 2009 (UTC)

Update 17:20, 3 June 2009 (UTC) for 0.8.0c, version from debian sid (unstable). gnunet-search gpl worked quickly for me; gnunet-download with the URI did nothing for about 5 minutes, i killed it, and the same command a second time got the file within a few seconds or so. The daemon gnunetd is being happily used by numerous peers without any problem. But the gnunetd.conf file lacks some basic info on what the various parameters mean, e.g. is BPS bits/second or bytes/second? is QUOTA in bytes, kbytes, Mbytes, kibibytes, Mebibytes, or whatever? i didn't try the GUI. As for the subject of this talk page section: the package was for some reason not included in the present debian stable (lenny), but it was in the previous stable (etch) and 0.8.0b is in debian testing (squeeze), so that means that in about 18-24 months (17:20, 3 June 2009 (UTC) + 18-24 months) 0.8.0+ will automatically make it back to the next debian stable version unless there are some special concerns. So it's still clearly in development mode, but usable by someone with moderate GNU/Linux sysadmin skills with minimal effort. i would guess (wild speculation) that lack of even one-line documentation in the config file for the daemon might have stopped it getting into lenny.

Anyway, for wikipedia purposes, in principle we should have a report by some external website, not just wikipedians. Boud (talk) 17:20, 3 June 2009 (UTC)

attacks on anonymity

there should be a section on this —Preceding unsigned comment added by 201.152.49.159 (talk) 20:20, 9 May 2008 (UTC)

No Windows Binaries for download

On https://gnunet.org/downloads there are no Windows binaries for download, so listing that Windows is supported is not accurate. I have not yet tested if the sources compile on Windows, but even if they did, that doesn't mean there is a windows version available, that means you can make your own unsupported windows version (there does not even appear to be windows specific build/install/config documentation on their website). BrainSlugs83 (talk) 04:16, 3 October 2012 (UTC)

Hmmm, looks like I'm mistaken on it not being supported. Glad I did some extra digging. Their documentation tree on the right side of the page confused me a little bit. The link is here: https://gnunet.org/w32-build BrainSlugs83 (talk) 04:34, 3 October 2012 (UTC)

Double hmmmm, looks like the build instructions are super out of date and most of the links to precompiled libraries / binaries are completely wrong -- I can probably still get it compiled, but it's a bit like yak-shaving and a lot of trial and error -- I'm going to have to go back to my original statement that Windows is clearly not supported. :-/ BrainSlugs83 (talk) 07:41, 3 October 2012 (UTC)

(Sentence before this: I'm one of the GNUnet developers. What I write here is less polished than what ends up in the repository) -- Sorry for taking this long to find this. 1.5 years ago I started reworking the documentation and all the associated instructions. I wonder how OS X and Windows ended up in the wikipedia article for our software? Officially we do not support them. Our scope of support for operating systems (I will commit this this weekend to the README file) is Free Software Operating Systems. We take patches for Windows and OS X and other proprietary OS if they do not break any other system support. support for those OS is up to the OS developers or users. Regression can happen with progress we make, but we won't break support with malicious intent. From the build-system side, native cross-compiling could be achieved and implemented (in addition to the rest of the codebase) but someone has to do the work. FreeBSD removed it a while back for technical reasons I want to look into this year, it should still work on their OS. If the statement in the Box of the page 'Operating Systems' is meant to state the *officially* supported Operating Systems, I would suggest with the GNUnet hat on that you remove 'OS X' and 'Windows' from that list. -- regards ng0/N. Gillmann / GNUnet (Ngzero (talk) 13:21, 6 June 2018 (UTC))

I haven't removed 'OS X' and 'Windows', but instead clarified the split you've described with as less words as possible in the infobox. That's a more precise description of what's actually the real deal. A development team can take up this position, like you described. That's completely okay. But also relevant: I consider it to be important, that the article has a seperate section, where this operating system support policy is described in a more detailed way than in the infobox. Regarding the degree of detail a mix between the statment of Ngzero and the minimal version in the infobox seems to be appropriate. At least this suggested 'operating system support policy' section in the article should supplement the minimal version in the infobox by adding the point, that the development of the GNUnet is open to community contributions for other operating systems. Maybe something like:

'Primarily development effort is focused on Free Software operating systems, but the GNUnet development team is open to community contributions developping support of other operating systems.'

You've mentioned, that you've started reworking the documentation and all the associated instructions of the GNUnet project. In it's current form the wikipedia article seems to be a little bit outdated. Give us a list of all links to most recent and striking description of the GNUnet in form of text, pictures, and video, and we'll see what we can do about it. That would help for improving the article right now all around. For the long run, for maintaining the article, to have a similar link list here on this discussion page would be of special interest: A list of links, which will always have the most current and striking description of the GNUnet. That would enable other people to look up that places, compare if the article is still an accurate representation of the most current version of the GNUnet including future development tendencies, and update the wikipedia article, if necessary. It's of highest importance, that the english wikipedia article describes software projects with most current information and in the most striking way, because the english wikipedia article is a reference point for many other wikipedia articles of the same topic in other languages. — Preceding unsigned comment added by 178.201.144.69 (talk) 10:21, 10 June 2018 (UTC)

Out of Date/Incomplete Information on Java Port

It appears the latest version 0.9.3 adds a Java bindings package (i.e. so that Java developers can code against the GNUnet api). Not sure what that other java mention in the article was about, but it links directly back to this article with absolutely no explanation, which is ridiculous. BrainSlugs83 (talk) 04:19, 3 October 2012 (UTC)

GNS

It looks like there is something else new in this release: "This release contains the beginnings of a new GNUnet-based Naming System (GNS), implementing a fully decentralized, backwards-compatible replacement for DNS (many important features and documentation are still missing, but the foundations are there)" -- which is really cool and exciting, which makes me even more upset that there are no windows binaries available.

The documentation link for GNS is here: https://gnunet.org/gns BrainSlugs83 (talk) 04:16, 3 October 2012 (UTC)

overall update 2018/2019

The article is a bit outdated. So let's give it an overall update. Here's the requested reference list of all links to most recent and striking description of the GNUnet in form of text, pictures, and video.

• https://gnunet.org/ - title image
• https://gnunet.org/ - About
• https://gnunet.org/git/gnunet.git/tree/README - README of GNUnet
• https://docs.gnunet.org/ - GNUnet documentation, more relevant sections: 1 Preface, 1.3 Introduction, 2 Philosophy
• https://media.ccc.de/v/34c3-ChaosWest-6-privacy_oriented_distributed_networking_for_an_ethical_internet_including_50_subsystems_of_gnunet - video of recent talk 'Privacy-Oriented Distributed Networking for an Ethical Internet - including 50 subsystems of GNUnet' (2017-12-27)
• https://grothoff.org/christian/habil.pdf - habilitation about GNUnet
• https://gnunet.org/git/www.git/tree/architecture.html.j2 - Code for architecture section of the new gnunet.org website scheduled to launch in approximately 2019
• https://gnunet.org/git/presentations.git/ - project internal list collecting self-hold GNUnet related talks and presentations, well curated, meaning always up to date

All what needs to be done now is synthesize everything for a broad audience, section by section. — Preceding unsigned comment added by 88.152.172.49 (talk) 12:17, 30 July 2018 (UTC)