26

Re: And if all from the beginning?

Hello, elmal, you wrote: E> OS can be necessary that this microwave to control from a mobile phone. Thus the microwave program says to an exterior surrounding that she a microwave, is able to turn on, install capacity from a certain range, is able to twist a grill and to speak where it is allocated. Unless for this purpose OS is necessary to a microwave? And whether it is necessary to break and make a fresh start all, what there would be such functionality? In 1999 questions of communications of things it was offered to solve radio-frequency labels and the elementary protocols.

27

Re: And if all from the beginning?

Hello, Pavel Dvorkin, you wrote: PD> the Request nevertheless to remain within the limits of a subject. To this demiurge of possibility to change iron, especially an external world I did not give And in vain. Because the modern iron is that generally because of a load of compatibility with written before a software.

28

Re: And if all from the beginning?

Hello, vdimas, you  V> And in vain. Because the modern iron is that generally because of a load of compatibility with written before a software. Here 2 moments At first, architecture x86/x64 would suggest to alter everything, here the doubt is not present, there is nothing to consider. And which it is necessary - on this subject and without that arguings much Secondly, I do not quite agree with the thesis. Linux where only does not work, and appear the new processor - will be and to work on it. Windows too and everything that above OS - so it is absolutely feeble from the processor depends

29

Re: And if all from the beginning?

PD> That from this, what is put in existing a software, made how and it would be necessary "on mind" to make, and what it would be necessary to make differently, yes only, alas, it is impossible - hinder damned compatibility and huge operating time? Every time accurately to solve, for whom the software (beginning from operating systems) becomes user - multjashno-odnookonno-paltsetychnoe for those a coma to strain a brain painfully and terribly, or "engineering" for not having such pathology. If very much it is necessary to let out 2 different versions. But not to try to reduce these "2 worlds, 2 morals" to  to common denominator as do a microsoftware with  and on an occasion at them crowd of imitators.

30

Re: And if all from the beginning?

Hello, Pavel Dvorkin, you wrote: PD> What you made of the existing in the almost same type in what it exists now and what would make in another way? Would invent the Fortran 77, probably. But with reserved keywords and without . And any GOTO. Arithmetical IF - our all!

31

Re: And if all from the beginning?

Hello, Pavel Dvorkin, you wrote: PD> Iron remains. Without changes. What so to waste time on trifles that? Iron development came into cruel lockup, and recent (but for certain not last) vulnerability already hammer in a cover of a coffin nails. New iron is necessary!... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

32

Re: And if all from the beginning?

Hello, Pavel Dvorkin, you wrote: PD> What you made of the existing in the almost same type in what it exists now and what would make in another way? As already mentioned - would alter JavaScript. That is would leave its very similar in general, but would repair trifles - all these autocoercions in comparing and other WTF which now cannot throw out because of backwards compatibility. C on itself idea to have HTML5+JS for mobile applications - sensible. Trifles spoil all pattern. Would make SQL . Would refuse hardware insulation in OS, leaving only  with verification. Hardware protection - expensive and unsafe (see Meltdown). Static verification is much more reliable.

33

Re: And if all from the beginning?

Hello, Sinclair, you wrote: S> Hello, Pavel Dvorkin, you wrote: S> would Refuse hardware insulation in OS, leaving only  with verification. Hardware protection - expensive and unsafe (see Meltdown). Static verification is much more reliable. Will not be slowly? And provocative (knowing your relation to it) a question. HTTP would not change? Or, can, with something other replaced?

34

Re: And if all from the beginning?

Hello, Pavel Dvorkin, you wrote: PD> will not be slowly? Will be faster. Because there are no switchings between rings for syscall. PD> And provocative (knowing your relation to it) a question. PD> HTTP would not change? Or, can, with something other replaced? Here I do not know. HTTP2 quitted after I ceased to read RFC on protocols; therefore I do not know, what there are taken measures. As a whole, HTTP/1.1 it is already rather fine. I do not know that such it is necessary to do to rest against its restrictions in client server applications.

35

Re: And if all from the beginning?

Hello, Sinclair, you wrote: S> Hello, Pavel Dvorkin, you wrote: PD>> will not be slowly? S> will be faster. Because there are no switchings between rings for syscall. And protection? You suggest it to do at purely software level? But then and to check each time it is necessary with software usage, it hardly will be quickly char * p = to something *p = 0; And how here at software level to check, whether I have access right there? To build in on everyone mov check commands? PD>> and provocative (knowing your relation to it) a question. PD>> HTTP would not change? Or, can, with something other replaced? S> here I do not know. HTTP2 quitted after I ceased to read RFC on protocols; therefore I do not know, what there are taken measures. As a whole, HTTP/1.1 it is already rather fine. I do not know that such it is necessary to do to rest against its restrictions in client server applications. A question not in restrictions, and in architecture. It not seems to you, what the same interface GET with parameters it would be possible to replace with something more elegant? And at the same time Header to alter.

36

Re: And if all from the beginning?

Hello, Pavel Dvorkin, you wrote: PD>>> will not be slowly? S>> will be faster. Because there are no switchings between rings for syscall. PD> And protection? You suggest it to do at purely software level? But then and to check each time it is necessary with software usage, it hardly will be quickly PD> char * p = to something PD> *p = 0; PD> And how here on software level to check, whether I have access right there? To build in on everyone mov check commands? Obviously, access to the arbitrary address and arithmetics of pointers (and generally treatment of pointers as numbers) in the application-oriented code completely is forbidden. In system code it is possible to leave restricted possibilities of manipulation pointers safety of operations with which is guaranteed by the static verifier (but such - "char * p =" - anyway it is forbidden to something).

37

Re: And if all from the beginning?

Hello, AlexRK, you wrote: ARK> it is obvious, access to the arbitrary address and arithmetics of pointers (and generally treatment of pointers as numbers) in the application-oriented code completely is forbidden. In system code it is possible to leave restricted possibilities of manipulation pointers safety of operations with which is guaranteed by the static verifier (but such - "char * p =" - anyway it is forbidden to something). How you represent it to yourself? It is possible not to use, of course, pointers in Java, # or JS - a code where they and are not present. So after all in jvm.net most or the browser you will not forbid! If at all to forbid pointers (machine addresses) in  - it is necessary for each command working with storage, to switch in a kernel mode. It will be fantastic a procorf of speed. And to transfer jvm etc. to a kernel mode is it will be simple to mean transfer of all operation to a kernel. After that about safety of a kernel will speak difficult.

38

Re: And if all from the beginning?

Hello, CoderMonkey, you wrote: CM> That so to waste time on trifles that? Iron development came into cruel lockup, and recent (but for certain not last) vulnerability already hammer in a cover of a coffin nails. CM> new iron is necessary! Itanium.

39

Re: And if all from the beginning?

Hello, Sinclair, you wrote: S> would Make SQL . What for? Than type dialects t-sql do not arrange?

40

Re: And if all from the beginning?

Hello, Pavel Dvorkin, you wrote: PD> It is possible not to use, of course, pointers in Java, # or JS - a code where they and are not present. So after all in jvm.net most or the browser you will not forbid! Why is not present? You will forbid. Simply to do everything it will be impossible (for example, simply to add to the pointer unit and to address to this address). PD> And to transfer jvm etc. to a kernel mode is it will be simple to mean transfer of all operation to a kernel. After that about safety of a kernel will speak difficult. Yes, speech about transfer of all to a kernel mode (actually, no "rings" already also will exist). Safety is provided with static verification (as in Singularity or Rust).

41

Re: And if all from the beginning?

Hello, Pavel Dvorkin, you wrote: PD> And to transfer jvm etc. to a kernel mode is it will be simple to mean transfer of all operation to a kernel. After that about safety of a kernel will speak difficult. Though on a condition about a software here all forgot, iron remains such as was. And about iron, in particular, storage protection, here spoke in far 2005. Whether storage protection is necessary Oberons-wasps?

42

Re: And if all from the beginning?

Hello, AlexRK, you wrote: ARK> Yes, speech about transfer of all to a kernel mode (actually, no "rings" already also will exist). So, give in sequence. Rings will not be. Applications can use only managed code. Then a question - and who has the right to develop  the code? It a certain indirect vendor, or only the manufacturer of the processor and those to whom it it entrusts nevertheless can add?> safety is provided with static verification (as in Singularity or Rust). I understand, at what you drive. But Singularity did not go. Besides, I something am not assured that static verification is absolutely reliable.

43

Re: And if all from the beginning?

Hello, Privalov, you wrote: P> Though on a condition about a software here all forgot, iron remains such as was. And about iron, in particular, storage protection, here spoke in far 2005. About it not only spoke, it and was made (Itanium), yes only the architecture x86/x64 for itself managed to stand. Easier if to start it to consider -  turns to arguing of what should be the processor (because x86/x64 anybody will not protect architecture).

44

Re: And if all from the beginning?

Hello, Pavel Dvorkin, you wrote: PD> Applications can use only managed code. PD> then a question - and who has the right to develop  the code? It a certain indirect vendor, or only the manufacturer of the processor and those to whom it it entrusts nevertheless can add? Not is mandatory in applications there should be a managed code. Can quite be and statically verified  the code. Similar things for a long time already in development: https://en.wikipedia.org/wiki/Verve _ (operating_system) https://en.wikipedia.org/wiki/Typed_assembly_language Verve - statically verifiable OS, from top to bottom, it is absolute entirely - including the most low-level parts on the assembler (it, of course, not the simple assembler, and TAL - the assembler approaching for static verification). That is it is all  the code, but thus verified on a compilation stage. I think, in the future all to that and comes. Though for normal applications to write in controlled languages simply more cheaply, therefore, I think, the managed code too remains. PD> I understand, at what you drive. But Singularity did not go. Did not go, truly. But showed that similar implementation at desire is quite possible. (There will be such desire at someone or not and if will be when is all open questions.) PD> Besides, I something am not assured that static verification is absolutely reliable. On 100 %, of course, is not present. No less than processors are not reliable on 100 % (recent holes see).

45

Re: And if all from the beginning?

Hello, AlexRK, you wrote: ARK> On 100 %, of course, is not present. No less than processors are not reliable on 100 % (recent holes see). OK. This OS with verification is created. In it it is possible to write programs on  the code, but it verify. But verification not on 100 % is reliable. Moreover, I assume that it is much less reliable, than protection in the processor. There nevertheless from 0 rings to a kernel it is impossible to reach. Also it is not necessary to me about Meltdown and Spectre - all of them on record do not give access. Someone finds a method to bypass this verification. To bypass not because its algorithm contains errors that is why that it found a situation when verification does not work. Any artful exercises dementing the static verifier. Such can be? Nevertheless the program is executed in dynamics - you are assured, what static means it is possible to provide all cases? As a result it the master of the machine. Other protection is not present more. At it the full access to OS kernel. And if he is an author of a virus, that is does it meaningly - and what now to do?

46

Re: And if all from the beginning?

Hello, Pavel Dvorkin, you wrote: PD> But verification not on 100 % is reliable. Moreover, I assume that it is much less reliable, than protection in the processor. You never can tell. Hardly it is possible to measure it. PD> there nevertheless from 0 rings to a kernel it is impossible to reach. In the theory it is impossible, but who told, what in the processor there are no holes? Modern  - a piece very difficult. PD> Someone finds a method to bypass this verification. To bypass not because its algorithm contains errors that is why that it found a situation when verification does not work. Any artful exercises dementing the static verifier. PD> such can be? Quite can, I think. In practice, , the probability is not too great. But, of course, in this plan state something I cannot. PD> nevertheless the program is executed in dynamics - you are assured, what static means it is possible to provide all cases? It is theoretically possible to provide everything, in practice it is possible any hole and to pass. PD> As a result it the master of the machine. Other protection is not present more. At it the full access to OS kernel. PD> and if he is an author of a virus, that is does it meaningly - and what now to do? And now about the same. Bugs in the processor, bugs in a kernel, escalation of privileges - and at you the full access to OS kernel, you the master of the machine. Attacking the Windows Kernel Ring 0 to Ring-1 Attacks Now, of course, it is possible to tell that " defense" consists of two bastions - at first holes in a software, and then holes in gland. But what hinders to make some echelons in a software? I do not see the basic difference. UPD. By the way, here found about serious holes in , admitting breaking of structures of a kernel: https://www.theregister.co.uk/2015/08/1 … rocessors/

47

Re: And if all from the beginning?

Hello, Pavel Dvorkin, you wrote: PD> About it not only spoke, it and was made (Itanium), yes only the architecture x86/x64 for itself managed to stand. I to admit, on Itanium I know nothing. Tried to read, but something is not went. However, at me under the schedule labor fulfillments. It appears, they  influence ability quickly to digest the read. Briefly, it has a protection of storage or not? And if is, it is possible to invent a C ++ and if is not present, the Oberon? However, with last not all so is simple. More shortly, the Fortran , whatever one may do.

48

Re: And if all from the beginning?

Hello, Privalov, you wrote: P> I to admit, on Itanium I know nothing. Tried to read, but something is not went. However, at me under the schedule labor fulfillments. It appears, they  influence ability quickly to digest the read. P> briefly, it has a protection of storage or not? , did not understand. How there can be a processor of 21 centuries without storage protection? P> and if is it is possible to invent a C ++ and if is not present the Oberon? However, with last not all so is simple. More shortly, the Fortran , whatever one may do. A Fortran our youth

49

Re: And if all from the beginning?

Hello, AlexRK, you wrote: PD>> But verification not on 100 % is reliable. Moreover, I assume that it is much less reliable, than protection in the processor. ARK> there is no saying. Hardly it is possible to measure it. Perhaps I do not understand something, but int n, m; read (m, n);//input whence int a [m];//we select storage and we create an array a [n] = 1; How it is possible to check up this fragment statically? At correct m and n all will be remarkable. At big n we freely now can leave in kernel addresses, for as we receive AV And statically - as?

50

Re: And if all from the beginning?

Hello, lpd, you wrote: lpd> Hello, Pavel Dvorkin, you wrote: PD>> What you made of the existing in the almost same type in what it exists now and what would make in another way? lpd> I would replace With ++, Java and C# with one language without VM with garbage collection (). But about it different judgements much. Was such in Objective-C. Did not get accustomed. By the way, I do not know, why. Is , to throw light?