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=-1.9 required=5.0 tests=BAYES_00 autolearn=ham autolearn_force=no version=3.4.4 X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 103376,bc1361a952ec75ca X-Google-Attributes: gid103376,public X-Google-ArrivalTime: 2001-09-05 00:47:20 PST Path: archiver1.google.com!newsfeed.google.com!newsfeed.stanford.edu!news.tele.dk!small.news.tele.dk!130.133.1.3!fu-berlin.de!uni-berlin.de!ailinel.cbb-automation.DE!not-for-mail From: dmitry@elros.cbb-automation.de (Dmitry Kazakov) Newsgroups: comp.lang.ada Subject: Re: Progress on AdaOS Date: Wed, 05 Sep 2001 07:47:18 GMT Message-ID: <3b95d429.592218@news.cis.dfn.de> References: NNTP-Posting-Host: ailinel.cbb-automation.de (212.79.194.99) X-Trace: fu-berlin.de 999676039 5482800 212.79.194.99 (16 [77047]) X-Newsreader: Forte Free Agent 1.21/32.243 Xref: archiver1.google.com comp.lang.ada:12735 Date: 2001-09-05T07:47:18+00:00 List-Id: On Mon, 3 Sep 2001 19:16:11 +0000 (GMT), "M. A. Alves" wrote: >> >But why limit a file to an array of storage elements? Why not persistant >> >unbounded storage pool? Or persistant unbounded tagged classwide type? >> >Or persistant protected object? (Ada already has the concept of a >> >persistant storage mechanism, pragma shared_passive. Why not use that?) >> >> Amen. >> >> I would like to add, why to have files at all? There should be only >> objects allocated in one big virtual memory chunk. 64-bit address >> space allows to address every objects in every computer of the world. >> There should be no I/O, only memory mapping. > >Yes! With each memory cell classified according to: > > * readable or not > * writeable or not > * timeable access or not (for real time computations) > * persistency > * ... I think in an OO OS there should be more fine grained properties. For instance there should be different access policies regarding the object itself and its methods. For instance, an object created by a more privileged user [object] might have methods available for less privileged users [object]. Further, these methods can be executed either on the caller's or callee's [supervisor calls] contexts etc etc. >With the proper methods and exceptions e.g. Put, Get, Read_Error, etc. > >Some mixin, it appears. > >You still have _devices_. All cells in the same device would have the >same class(ification). In fact all there is is devices, not computers, as >a computer is a set of devices. So each device in the world would have an >id (some already do) and so each cell in the world has id (device_id, >cell_number) or so. But a 50/50 partition of the 64-bit addressing gives >c. 1 million devices with 1 million cells each: not enough. Go for >128-bit, 1 billion devices of 1 billion cells each. And I feel even that >is a bit tight. 256-bit? > >Now, "mapping". The OS would have device interfaces (libraries), and an >"overall" layer providing the universal mapping. Regards, Dmitry Kazakov