From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: X-Spam-Status: No, score=-1.4 required=5.0 tests=BAYES_00,URI_NOVOWEL autolearn=ham autolearn_force=no version=3.4.4 X-Google-Language: ENGLISH,ASCII X-Google-Thread: 103376,ddb34e4ee5e28db0 X-Google-Attributes: gid103376,public X-Google-ArrivalTime: 2004-04-27 02:08:34 PST Path: archiver1.google.com!news1.google.com!news.glorb.com!newsfeed.stueberl.de!news-FFM2.ecrc.net!news.iks-jena.de!not-for-mail From: Lutz Donnerhacke Newsgroups: comp.lang.ada Subject: Re: BIND Date: Tue, 27 Apr 2004 09:08:33 +0000 (UTC) Organization: IKS GmbH Jena Message-ID: References: <20040425224751.C907A4C4136@lovelace.ada-france.org> NNTP-Posting-Host: taranis.iks-jena.de Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Trace: branwen.iks-jena.de 1083056913 7954 217.17.192.37 (27 Apr 2004 09:08:33 GMT) X-Complaints-To: usenet@iks-jena.de NNTP-Posting-Date: Tue, 27 Apr 2004 09:08:33 +0000 (UTC) User-Agent: slrn/0.9.8.0 (Linux) Xref: archiver1.google.com comp.lang.ada:7527 Date: 2004-04-27T09:08:33+00:00 List-Id: * Alexei Polkhanov wrote: > On Mon, 26 Apr 2004 19:36:36 +0000, Lutz Donnerhacke wrote: >> First: Downsize the goals. > I would merge DHCP and DNS together in a single product. There is several > reasons for that. NO! It's dumb, dumb, and dumb again. >> Second: Implement a simple DNS proxy for NS, SOA, A, MX, CNAME and PTR. > Sounds almost trivial. Welcome to the real world. >> Third: Implement authoritive zones. > I suppose support for RDBMS and HUGE ZONES is what we need here. If you > have 2,000,000 record zone it takes 20 min to reload it. BIND has an > "rndc" tool which reloads zone files without restarting whole > thing. When you have such big zone you expect many changes to it too. rndc is a frontend to Bind, not a solution! > This problem needs some speacial approach. Yep, that's why leave it out as long as possible. >> Fourth: Extent the software to other types namely to IPv6. > IPv6 is would be a small implementation detail, don't make it look like a > major undertaking, AAAA records plus some IPv6/IPv4 migration stuff. Wrong! There is no working IPv6 implementation of GNAT.Sockets. There are years of failures in understanding and setting up IPv6. You are on the best way to redo all this mistakes again. > National languages support, special records and extentions which need to > be taken care of. Latest I've heard is an extention for IP Phones. GNA! Stop implementing anything! Please READ a specification first! You do not understand what you are talking about. Please drop this project! > "www.domain.com" notation is very technical, network - centric, so some > people trying to comeup with more human - readable names, and that going > to be in next versions of BIND. Why I cannot type "Congress Library" but I > have to go and type "dsgdsgf.dsfgdsfg.hfdhdf.gov" ? So most of us ill go > to Google because they cannot remember domain name and then will type it > in browser - why can't we skip this step ? You are looking f�r HDDB or other distributed hierarchical databases, but this is a major step away from the Bind project. (Please note, that I suggest as the very first step: "Downsize the goals!") >> Fifth: Rewrite the whole package from scratch to implement DNSSEC. > Yeah! Here is a Huge piece of code which is not the best part of BIND and > sounds like we already have some issues like encryption libraries. > Nothing I know of except thin bindings to OpenSSL. I can contribute even Hash, Symmetric, and Asymmetric Packages using generic class wide programming, but I doubt it will help anything. > Also there is TSIG, and SPLIT DNS (when you have all zone records divided > into subsets with different access rules or permissions). Right now this > feature "implemented" by config files hackery and having multiple > instances of BIND running on different machines sharing only parts of zone > records. Do not touch those areas. They are braindead klugdes. Do not implement such things. If you come to this requirement in 2007, you might consider views. >> I'd contribute: >> Stream_Text_IO (Basic Text_IO over Ada.Streams.Root_Stream_Type'Class) >> Skip_Stream and Memory_Stream (indexable Stream_Type over Storage_Elements) >> UDP and TCP server handling packages > > Would be nice to have this thing working on MOST unix platforms, so sounds > like Florist can be used here... So please stick to Florist. > Means it also nice to find a _sponsor_ at least for some UNIX hardware and > some network for testbed... GNA? From what system do post? Ada is for portable programming! You do not need any other hardware than a low end development system. (My programming, testing, and debugging enviroment is a Pentium 133Mhz with 32MB of RAM for Unix/Linux development.)