Discussing kernels in the second para could intimidate some users, and somewhat reinforces stereotypes about Linux users :-). That could be removed, as it is discussed later on.
There are significant inconsistencies with the refs, and they don't all use {{cite web}}, which will help fix that. Some say The Debian Project, others are debian.org. Dates are also not consistent, and some titles are missing. A few links are also dead.
Done
Mentioned inconsistencies are fixed.[1][2] I hope the remaining inconsistencies are not significant. Please indicate if reference consistency should be improved further. 84.127.80.114 (talk) 22:21, 25 May 2014 (UTC)Reply
144 appears to be out of date, and a {{tl|as of}] could be employed here.
GLAN Tank support is already mentioned in the next reference. As I understand, this Japanese reference was meant to be a wiki link for more information, but GLAN Tank does not have an article in the English Wikipedia. 84.127.80.114 (talk) 22:40, 28 May 2014 (UTC)Reply
"suggestions point to aptitude..." is a bit weasley, I also can't see much about package metadata in the source, and don't see why that is a big thing.
As a matter of fact, the source says: In case of doubt, please use the apt-get and apt-cache commands over the aptitude command. Search on metadata is just a feature. 84.127.80.114 (talk) 19:22, 29 May 2014 (UTC)Reply
APT tools for online repositories - the content under this heading repeats some of what is said earlier.
I have installed Debian wheezy on a virtual machine to verify the claim on amd64. The machine has 60 MiB of RAM and 1 GiB of disk space, without network. I used the CD 1 from the set. With a 128 MiB swap partition, the system has around 524 MiB available. I can use the nano text editor and play with aptitude's minesweeper.
The key in the installation process is to create the swap partition and activate it manually. This procedure is not trivial.
There are reports about previous versions being usable,[4] but I cannot find sources about wheezy. Anyway, the article does not claim whether the systems are usable or not. Since usable is relative, we leave that conclusion to the reader. 84.127.80.114 (talk) 18:43, 30 May 2014 (UTC)Reply
It is partner article (with editorial overview of IW this should not be a problem), old and does not tell the details about hardware in question, but the whole article is basically written to prove the statement, so it might be used for verification of the claim. — Dmitrij D. Czarkoff (talk•track) 21:22, 30 May 2014 (UTC)Reply
"must be added" - this makes it sound like 1Ghz + 200Mhz = 1.2Ghz, or something like that. "must be considered as well" is less misleading.
There are quite a few refs that appear to be slightly misused. For example, "Debian is known for its serious manifesto..." is referenced to https://www.debian.org/social_contract, which only states that the manifesto exsists, not that it is known for it or that it is serious. The root of the issue is probably that so many refs go to debian.org, which is great for some things but not so great for others. There are no links to The Register or Ars Technica, both of which are great starting points for this kind of thing.
"Each of them sustains some..." - I'm not sure if this sentence is necessary. I think more explanation about the actual development process would be helpful, especially how it differs from a commercial process.
"Informally, Internet Relay Chat channels..." - this kind of detail could be trimmed, but if a decent source can be found it would make an interesting addition that says something about the devs.
More secondary sources are needed around this point - phrases like "steady influx of applicants wishing to become developers" need secondary sources to be properly credible, as do descriptors such as "elaborate."
"matured and the goals for the next release are met" - "are complete and have been sufficiently tested", or something to that effect, makes it sound more natural.
The paragraphs on security need more secondary sources for obvious reasons. The Reception section at the end should be broken up, and the bit about the RNG added in here.
The Reception section, as per above, should be merged into the rest of the article. Server Distribution of the Year from 2007 can probably be removed somewhere along the way.
First pass done. I'll let you make some fixes and then go through it again. Thanks for being so quick at responding to points. Jamesx1234522:53, 3 June 2014 (UTC)Reply
The intro still isn't "right." It needs to be a bit longer, and maybe less technical. Ideally, any refs would bee redundant to the body of the article and could be removed.
"The Debian Installer team announced that the first CD includes GNOME thanks to their efforts to minimize the amount of disc space GNOME takes up." - this is probably longer than it has to be.
Debian version 5.0 "Lenny" was the first official Live CD release. - I removed this as I don't think it is terribly useful, as every feature that wasn't in 1.0 had to be added somewhere.
The section about Packages is in general short of a few refs. In particular, there are no refs between "An APT tool allows administration of..." and "...include Software Center, Synaptic and Apper."
On the Debian website, repositories and distros seem to be stylised as sid, stable etc (both the "state" and its alias.) There is quite a bit of inconsistency around this, some of it added by me.
"The arm (OABI..." - arm has not been mentioned before, but armel has. There is also some unclear formatting here, and you might want to double check everything that is said or implied.
The problem is that the expression "developed by volunteers" misleads to think "Debian developers". There is no definition of a "Debian volunteer". 3000 developers is plain wrong, even with Debian maintainers. 3000 volunteers only makes sense with a broader meaning, including patch submitters, users helping in mailing lists and conferences, etc. 84.127.80.114 (talk) 10:40, 21 June 2014 (UTC)Reply
"unstable (also known as "sid"), - here, distros are stylised as "sid" rather than sid.
The vulnerability was serious and there was a negative reaction, but I have not found much criticism. Quoting Tim Hudson from OpenSSL, "Attributing blame for this issue is a pretty pointless exercise IMHO." 84.127.80.114 (talk) 17:17, 21 June 2014 (UTC)Reply
Is 7 or 7.5 going to be used here? It should be harmonised with the version given in the infobox.