276

Re: And if all from the beginning?

Hello, netch80, you wrote: N> This roughly incorrect comparing: N> 1. At Go simple syntax (here it it is strong in plus, they bypassed a standard rake the C-C-company), and all possibilities which compilers of any Cs ++ brake are removed. Well we speak about a stage, which already after parse. N> 2. At Go very feeble optimization - it is killed as a heap of rules of type "expression always arithmetics of integer numbers on the unit" is calculated from left to right ",", "function arguments  are calculated from left to right", etc., and the execution organization - for example, now internal ABI demands to save by a call of other function all intermediate data on a stack (callee-saved registers are not present generally, short of auxiliary like SP. Therefore it plainly is not spent for it. It yes. But here in my opinion business not only in rules, and and that in respect of optimization anybody while especially also was not strained, so the optimizer there naive enough. It, truth, it is easy and without being strained is able to do some features which in Si it is impossible to make or it is very difficult. For example,   and escape analyser works even for calls from one packet in another. And  at it always smart if function is not necessary, it nobody . It here advantage of that at it not #include, and import, and the compiler always sees that is imported, and not just prototypes of functions. Besides, at it the compiler is able to select  constructions in the text, and to compile in special way. It, generally, in my opinion, rich thought instead of spending life for creation of the high-grade static analyzer, simply to learn to select standard constructions which all use, and to invent, how they can be optimized. However, while it of such constructions selects a little. But in general, with your statement that at it optimization nasty, I will agree. N> 3. Projects of the size Firefox at it does not exist. As the compiler leads on it, whether dies under loading - as far as I know, nobody checked. And here I just assume a maximum of a rake. For example, implicit inheritance of a class to the interface only upon compatibility of methods - that at hundreds thousand classes can turn to a nightmare by correspondence search. And he also should not search anything. It has value of specific type, and the interface type to which it should be transformed. It should check up compatibility only. Other question that these here  from specific type to the interface, I do not know, how they at it are called, he tries not to double them, therefore he should be convinced that such still is not present. I, by the way, am not assured that it does it statically. It precisely absolutely should do it in  (for example if we transform type T at first to interface I1, and then in its subset, interface I2  will be I2-> T, instead of I2-> I1-> T, and it will be clarified only in ). But whether it here does it statically there where it is possible, I simply at all do not know Pzz>> As far as I understand, at go  does approximately that here suggest to do at setting. N> I something did not see, that at it differed the compiler and  as different entities. In 99.9 % of cases simply throw it to compile all dump of source codes together. The compiler does not do anything essentially artful (see above). The Intermediate language too at it is not present, all goes at once to the assembler. Also what here the general? Differ, differ, still as differ. At it the compiler leaves on itself files.a, with pair mysterious files inside. And as far as I understand, not the code, and result of parse (with all checks, it is finite) there lies. And if in the project the same packet is used some times, it only once compile. And here already  screws together all it together. Note that   is able to decompile already screwed together executed file, instead of these intermediate nepojmi-that-for-feature.

277

Re: And if all from the beginning?

Hello, lpd, you wrote: lpd> Because of problems of authors of a software under Windows I suffer even less, than because of problems . I not against paid closed-source, however upon it appear unnecessary because of a considerable quantity open-source a software. It personally to you it appears is not necessary. And here remaining suddenly it appears all the paid software with the closed source codes is necessary. Well, that is I not against  exist free open-source, but it cannot force out a paid software from appropriate niches. And I do not see perspective that it once could. If to look at successes of implementation of new platforms for the last decades suddenly it appears that the success directly is connected to monetization possibilities. iOS flied up not because in it there were any progressive ideas that is why that APPL constructed a development ecosystem so that simple boys could earn on this platform. And at once - motivation, a competition, perfection, progress. And if I let out tomorrow the free OS calculated for a free software developers since in it for it do not pay bluntly do not go to it. And, hence, users because there there is no software do not go also. Well and all - the problem is solved; audiences are not present, the project is closed.

278

Re: And if all from the beginning?

Hello, Sharov, you wrote: S> And  because of typification there where it is not necessary . That??? S>>> And in remaining sql quite the typed language of requests -  (union) hens and eggs does not turn out. S>> yes as  to do. You look at the table implicit conversions in T-SQL. S> I about ansi sql, instead of dialects. I for you am not in time - just was T-Sql Anything, ANSI is even worse.

279

Re: And if all from the beginning?

Hello, scf, you wrote: scf> Hello, Pavel Dvorkin, you wrote: PD>> Hm. Generally it is - the MAC-address. It is almost beaten. But for limits  it is not transferred. It is necessary? What for? scf> That after moving to other country you include a coffee maker, it climbs on the server. And the server knows that it is the same coffee maker it now in other country. Solves an authentification and localization problem. Actually instead of "address" here it is necessary  in each pan of steam of keys - public/private. Well, as in the modern IP-phones. It is surprising, as vendors in this branch  15 with implementations absolutely full of holes like "we do request on FTP and in clear we swing a file of adjustments for the device, together with user name/password on SIP the server". Three generations of crutches it was replaced, while they not  to use SSL and PKI.

280

Re: And if all from the beginning?

Hello, Sinclair, you wrote: S> And if I let out tomorrow the free OS calculated for a free software developers since in it for it do not pay bluntly do not go to it. And, hence, users because there there is no software do not go also. That you speak, is true for some niches: for example, open source games, it is valid, less, than paid. To a lesser degree it is true and for applications with the graphic interface. Open source developers generally write that is interesting by him and it is necessary. The software ' for programmers and administrators ' Thus turns out that appeared  on servers and at workstations of programmers. As a result on servers open source the software forced out the commercial. Being returned to a subject native vs VM: I Will agree that at a great number of processor architecture to spread programs in the intermediate codes it makes sense. On the other hand, the favor of a zoo of processor architecture is unobvious is if only complicated life to programmers since as a matter of fact these architecture a little than would differ.

281

Re: And if all from the beginning?

Hello, lpd, you wrote: lpd> Being returned to a subject native vs VM: lpd> I Will agree that at a great number of processor architecture to spread programs in the intermediate codes it makes sense. lpd> on the other hand, the favor of a zoo of processor architecture is unobvious is if only complicated life to programmers since as a matter of fact these architecture a little than would differ. Well so if architecture particulars are visible only to system programmers - to authors AOT/JIT of generators - that that before to remaining programmers? And about complication there are some doubts - for example, the same Intel easy would throw out FPU which became outdated at once on several parameters, and it would become easier hardly less, than all.

282

Re: And if all from the beginning?

Hello, netch80, you wrote: N> Well so if architecture particulars are visible only to system programmers - to authors AOT/JIT of generators - that that before to remaining programmers? If to compile programs from the intermediate code in the processor codes right after installations you are right. And a JIT compilation it is not absolutely transparent any more since complicates process of start of the program. N> and about complication there are some doubts - for example, the same Intel easy would throw out FPU which became outdated at once on several parameters, and it would become easier hardly less, than all. It would be required to standardize the intermediate code and its dynamic  (if it to leave). It would be desirable, that all it has been justified by substantial improvements of processors, instead of is simple switch-off of support of legacy-instructions. The processors which are let out by different firms on one architecture (Intel and especially Arm) and without that differ among themselves, remaining compatible. It is not eliminated that it is enough of it for development of technologies of processors. It earlier at all was Windows, and for it shareware programs which, really, someone could not recompile under new architecture. Here it is present legacy. And now: - office users the restricted circle of the programs is necessary, which vendors  under the new processor, and are not present at office users of need for new processors faster 3. - the house user the browser and the videocard with support of video and 3d schedules are necessary only. - on servers at all Linux and programs from open-source repositories, which quickly  community. It turns out, the incompatibility problem remained in 90 when new processors appeared more often.

283

Re: And if all from the beginning?

Hello, Sinclair, you wrote: S> Hello, Sharov, you wrote: S>> And  because of typification there where it is not necessary . S> That??? I meant that typification is not free. Also would lead to complication ansi sql which would be necessary  a miscellaneous . S> I for you am not in time - just was T-Sql S> Anything, ANSI is even worse. I initially spoke about sql (ansi) and its dialects of type t-sql. It quite suits the tasks, and minimum typification is present.

284

Re: And if all from the beginning?

Hello, WolfHound, you wrote: WH> Well here nobody hinders actually such microprograms to be statically typified. WH> moreover to write them it is possible not in a notepad, and in high-grade  which if to do it will be correctly launched for fractions of a second. WH> Sublime it is launched instantly. When  there is  it too will start fractions of a second. Now at start. brakes. In the theory you like the rights... But we look at practice (on languages existing at present and tools) and we see that all similar tasks dare exceptional on dynamic language. Something can we do not consider in our reasonings?))) WH>>> 2) console application creation on C# occupies some seconds. _>> Ugu, in imaginations.) WH> In practice. Well give we look at a reality. Here a real example from life about which I wrote above. The script means should for everyone fb2 a file in the current directory to fulfill conversion of its contents (to transform all html entities to their junikod-representation) and to thrust a file in zip archive with a name name.fb2.zip. Plus to it the source file should be just in case saved under a name name.fb2.bak (that analysis programs fb2 files it any more did not see). Implementing all is the Python a script occupied from me exactly 9 lines of the laconic code which I wrote the order 3-yoh minutes from which one spent for search in Google of function name and the unit from standard library for conversion html entities. Start Notepad ++ and then a script naturally instant. And for what time you from zero will write on C# the similar decision and how many lines of the code it will occupy? _>> the Problem of such systems (by the way Midori/Singularity it is far not the first in this area, was for example such http://www4.cs.fau.de/Projects/JX/index.html the decision much earlier) that they very categorical: or all application-oriented a software is written in appropriate controlled language or the system simply does not work (it is impossible to write one critical slice on the bare assembler). Well and accordingly while I will not see implementation for example the codec h.264 in the controlled language, not conceding on efficiency classical  to implementations in the normal future of similar OS I will not check. WH> with it problems are not present. WH> Dafna can prove any imperative code including the assembler. Well I can already check in such variant, but it will be already rather far from Mirodi and to it similar... By the way, and somewhere it is possible to look on: - examples of application of Dafny in any real projects (it is not important in what languages) - though one example for other languages (start up and not such exotic, how the assembler)?

285

Re: And if all from the beginning?

Hello, Sinclair, you wrote: _>> So I as though quite approve idea of propagation  (well naturally not such as in.net, and it is faster such as llvm, but not an essence). Only here similar propagation of a software is not connected to the fact and verification possibility in any way. And here is how time into the account of possibility of such verification at me is doubts. Truth so to say theoretical (I somehow yet did not write any similar verifier). S> Not so understood, in what of doubt. "Such verification" to itself is quite built in in quite itself real-life systems. There are public examples?)

286

Re: And if all from the beginning?

Hello, alex_public, you wrote: _> In the theory you like the rights... But we look at practice (on languages existing at present and tools) and we see that all similar tasks dare exceptional on dynamic language. Something can we do not consider in our reasonings?) )) the dynamically typed language Interpreter can write . The Statically-typed language with an output of types can write far not everyone. And the majority of those who can the scientists torn off from a reality at which always turns out . And remaining the big money pays for other operation. _> well give we look at a reality. Here a real example from life about which I wrote above. The script means should for everyone fb2 a file in the current directory to fulfill conversion of its contents (to transform all html entities to their junikod-representation) and to thrust a file in zip archive with a name name.fb2.zip. Plus to it the source file should be just in case saved under a name name.fb2.bak (that analysis programs fb2 files it any more did not see). Implementing all is the Python a script occupied from me exactly 9 lines of the laconic code which I wrote the order 3-yoh minutes from which one spent for search in Google of function name and the unit from standard library for conversion html entities. Start Notepad ++ and then a script naturally instant. And for what time you from zero will write on C# the similar decision and how many lines of the code it will occupy? If in C# there will be necessary libraries all will be approximately as. _> Well I can already check in such variant, but it will be already rather far from Mirodi and to it similar... By the way, and somewhere it is possible to look on: Mirodi this continuation of operation on subject Singularity and Verve. Because of their complexity, a holy grail of software verification has been to verify properties of operating systems. Operating systems are usually written in low-level languages, such as a C, that provide very few guarantees. The Singularity project took the approach of writing an operating system in C#, a type-safe, memory-safe language. A weakness of this approach is that operating systems necessarily need to call lower-level code to, for instance, move the stack pointer. Verve addresses this problem by partitioning the operating system into verified assembly that is required to be low-level and a trusted interface to rest of the operating system, written in C#. There is a trusted specification that guarantees the low-level assembly code does not mess with the heap and that the high-level C# code does not mess with the stacks. In the given approach nothing forbids to have the user verified assembly code. Moreover I doubt that it generally is necessary at presence halide. _> - examples of application of Dafny in any real projects (it is not important in what languages) was not interested. 1) Dafna imperative  language. Accordingly to stretch it on any of popular languages a trick. 2) Dafna analyzes each function irrespective of others. So the analysis it is possible  and result . Those there will be no problems even in terabyte of source codes. _> - though one example for other languages (start up and not such exotic, how the assembler)? And what difference what to prove the imperative code which changes local variables or the assembly code which changes registers?... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

287

Re: And if all from the beginning?

Hello, lpd, you wrote: N>> Well so if architecture particulars are visible only to system programmers - to authors AOT/JIT of generators - that that before to remaining programmers? lpd> if to compile programs from the intermediate code in the processor codes right after installations you are right. And a JIT compilation it is not absolutely transparent any more since complicates process of start of the program. You also will not note this complication. Delivered a product in a type  IL, data files, etc. - and start  on any executes RTLD. It and now everywhere nontrivial enough - reading ELF/COFF/etc., determination of segments for loading, calculation of relocation,  autoconnected SO/DLL - if to it will be added conversion read IL the code in commands of the local processor, it only decelerates start. A transparency for the author of the program will be saved on the full. N>> And about complication there are some doubts - for example, the same Intel easy would throw out FPU which became outdated at once on several parameters, and it would become easier hardly less, than all. lpd> It would be required to standardize the intermediate code and its dynamic  (if it to leave). It would be desirable, that all it has been justified by substantial improvements of processors, instead of is simple switch-off of support of legacy-instructions. Well examples Java/.NET/etc. Show that advantage in the form of notorious "write once, run anywhere"  is - and where it is not present, a problem almost always not in the processor. Exceptions go on type cases "we here anything vectorial we can not get => FPS <= 10". lpd> the Processors which are let out by different firms on one architecture (Intels and especially Arm) and without that differ among themselves, remaining compatible. It is not eliminated that it is enough of it for development of technologies of processors. lpd> It earlier at all was Windows, and for it shareware programs which, really, someone could not recompile under new architecture. Here it is present legacy. lpd> And now: lpd> - office users the restricted circle of the programs is necessary, which vendors  under the new processor, and are not present at office users of need for new processors faster 3. Even there is so, office it not all, and optionally most monetary segment. lpd> - the house user the browser and the videocard with support of video and 3d schedules are necessary only. And vendors of games eternally do not have speed. lpd> - on servers at all Linux and programs from open-source repositories, which quickly  community. Too not at all... On last operation we had many acquaintances with HPC packets of calculation something artful, for which their authors shivered, even leasing on the servers adjusted by them. But I agree that it  a case. lpd> It turns out, the incompatibility problem remained in 90 when new processors appeared more often. Yes, partially it is solved. But it is far not completely.

288

Re: And if all from the beginning?

Hello, WolfHound, you wrote: _>> In the theory you like the rights... But we look at practice (on languages existing at present and tools) and we see that all similar tasks dare exceptional on dynamic language. Something can we do not consider in our reasonings?))) WH> the dynamically typed language interpreter can write . WH> the Statically-typed language with an output of types can write far not everyone. WH> and the majority of those who can the scientists torn off from a reality at which always turns out . WH> And remaining the big money pays for other operation. I.e. you want to tell, what is simple while nobody mastered to make the correct statically-typed language approaching under  style of usage?) _>> Well give we look at a reality. Here a real example from life about which I wrote above. The script means should for everyone fb2 a file in the current directory to fulfill conversion of its contents (to transform all html entities to their junikod-representation) and to thrust a file in zip archive with a name name.fb2.zip. Plus to it the source file should be just in case saved under a name name.fb2.bak (that analysis programs fb2 files it any more did not see). Implementing all is the Python a script occupied from me exactly 9 lines of the laconic code which I wrote the order 3-yoh minutes from which one spent for search in Google of function name and the unit from standard library for conversion html entities. Start Notepad ++ and then a script naturally instant. And for what time you from zero will write on C# the similar decision and how many lines of the code it will occupy? WH> if in C# there will be necessary libraries all will be approximately as. All  are available (anyway judging by the documentation): https://msdn.microsoft.com/ru-ru/library/7c5fyk1k (v=vs.110).aspx and https://msdn.microsoft.com/ru-ru/librar … on.zipfile (v=vs.110).aspx. But thus I all the same extremely doubt in 9 line and 2-uh minutes on C#. _>> Well I can already check in such variant, but it will be already rather far from Mirodi and to it similar... By the way, and somewhere it is possible to look on: WH> Mirodi this continuation of operation on subject Singularity and Verve. WH> In the given approach nothing forbids to have the user verified assembly code. I in course. But Mirodi and Singularity is besides not real projects. In that sense that not used somewhere in real operation. WH> moreover I doubt that it generally is necessary at presence halide. Well SIMD is not unique application of the low-level code. For example a spelling any VM in language of type Java or safe C# seems to the extremely doubtful. _>> - examples of application of Dafny in any real projects (it is not important in what languages) WH> was not interested. WH> 1) Dafna imperative  language. Accordingly to stretch it on any of popular languages a trick. WH> 2) Dafna analyzes each function irrespective of others. So the analysis it is possible  and result . WH> Those there will be no problems even in terabyte of source codes. _>> - though one example for other languages (start up and not such exotic, how the assembler)? WH> And what difference what to prove the imperative code which changes local variables or the assembly code which changes registers? Me specific real examples of usage here interest. I an expert, instead of the theorist.)

289

Re: And if all from the beginning?

Hello, netch80, you wrote: N> Hello, lpd, you wrote: N>>> Well so if architecture particulars are visible only to system programmers - to authors AOT/JIT of generators - that that before to remaining programmers? lpd>> if to compile programs from the intermediate code in the processor codes right after installations you are right. And a JIT compilation it is not absolutely transparent any more since complicates process of start of the program. N> you also will not note this complication. Delivered a product in a type  IL, data files, etc. - and start  on any executes RTLD. I most of all do not like debugging at a JIT compilation. In  the code there is only a source code and the binary code, and the second is easy for comparing with the first. At a JIT compilation the intermediate code is added, and any more so simply to understand that to what without the aid of VM, simply looking and  at storage of process. Like as Linux therefore it is pleasant to programmers that it can be disassembled on components and to debug or replace them. The intermediate code postpones low-level operations from the user. Though at writing of high-level programs the difference is not so considerable. lpd>> It turns out, the incompatibility problem remained in 90 when new processors appeared more often. N> yes, partially it is solved. But it is far not completely. Processors reached good level (if not a maximum), and the further magnification of speed is necessary only on servers, for development process (engineers and designers) and for games + for javascript. And games all the same optimize vectorial instructions. It turns out, the intermediate code solves only  problems (a software for designers/designers). I consider a question from positions Linux, and for open source the intermediate code is necessary as the machine the fifth wheel solving hypothetical enough problems . Here with compilation of the intermediate code in binary during the moment application preferences I would not argue, if a variety of processors led to necessity of propagation of programs in such a way.

290

Re: And if all from the beginning?

291

Re: And if all from the beginning?

Hello, lpd, you wrote: lpd> I most of all do not like debugging at a JIT compilation. In  the code there is only a source code and the binary code, and the second is easy for comparing with the first. At a JIT compilation the intermediate code is added, and any more so simply to understand that to what without the aid of VM, simply looking and  at storage of process. What for such complexities? lpd> like as Linux therefore it is pleasant to programmers that it can be disassembled on components and to debug or replace them. The intermediate code postpones low-level operations from the user. Though at writing of high-level programs the difference is not so considerable. Once again. All compilers have in themselves the intermediate code. To easier you it do not show. lpd> here with compilation of the intermediate code in binary during the moment application preferences I would not argue, if a variety of processors led to necessity of propagation of programs in such a way. So here about this variant also speak. About that that the intermediate code can work only through JIT you to itself . And a variety of processors is hindered only by absence of a software under new architecture.... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

292

Re: And if all from the beginning?

Hello, WolfHound, you wrote: WH> And to a variety of processors absence of a software under new architecture hinders only. It is necessary to explain only, different on +-10 % of the speed, however incompatible processors what for are necessary, or to prove the considerable lag Intel/Arm from theoretically possible processors. Cutting out of legacy-instructions from Intel about which spoke netch80, does not look the sufficient reason for adding of the intermediate code. I, of course, understand that the monopolism complicates technical progress, however it would be desirable to look at least at a prototype of the new super-processor before to change an infrastructure of installation of applications on billion computers.

293

Re: And if all from the beginning?

Hello, lpd, you wrote: lpd> It is necessary to explain only, different on +-10 % of the speed, however incompatible processors That generally so it excites you what for are necessary? The processor command system will worry only the vendor of the processor. And it is more than anybody. lpd> or to prove the considerable lag Intel/Arm from theoretically possible processors. And we do not know. The matter is that development of the processor business not the cheap. And if it cannot be let out on the mass market it does not pay off. And without its software on the mass market not to let out. And if all software is beaten by nails to one architecture... Also is not present.  it is not enough. It generally is a microscopic part of all software. lpd> cutting out of legacy-instructions from Intel about which spoke netch80, does not look the sufficient reason for adding of the intermediate code. There are many other reasons. 1) Reliability and safety of a software. The verifiable code  is much more difficult. 2) convenience of developers. If to have the intermediate code that it is not necessary to think of that that there for the processor. It is not necessary to collect a heap of versions of a software. The user does not need to think what of installers to download. 3) faster and sympathetic programs. When we do not need to be afraid of a multithreading for the compiler checks that there all is well possible very aggressively all . 4) the Considerable simplification of programs. With ++ it is a hell. And thus that I very well know it With ++. But except it I know more many that. And consequently I can compare. 5) release of new versions  with the new optimizer accelerates all programs. And not just those which developers were not too lazy to recompile. 6) possibility to create pair the processor and  the optimizer under it. What gain of productivity gives it there is no saying. Can be and in times. lpd> I, of course, understand that the monopolism complicates technical progress, however it would be desirable to look at least at a prototype of the new super-processor before to change an infrastructure of installation of applications on billion computers. We here consider a situation "all software disappeared". It is necessary to write from a blank leaf.... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

294

Re: And if all from the beginning?

Hello, WolfHound, you wrote: WH> Also is not present.  it is not enough. It generally is a microscopic part of all software. I hear words burned . I will not overpersuade, since it is normally useless. lpd>> cutting out of legacy-instructions from Intel about which spoke netch80, does not look the sufficient reason for adding of the intermediate code. WH> there are many other reasons. WH> 1) Reliability and safety of a software. The verifiable code  is much more difficult. WH> 2) convenience of developers. If to have the intermediate code that it is not necessary to think of that that there for the processor. WH> it is not necessary to collect a heap of versions of a software. WH> the user does not need to think what of installers to download. WH> 3) faster and sympathetic programs. When we do not need to be afraid of a multithreading for the compiler checks that there all is well possible very aggressively all . WH> 4) the Considerable simplification of programs. With ++ it is a hell. And thus that I very well know it With ++. But except it I know more many that. And consequently I can compare. WH> 5) release of new versions  with the new optimizer accelerates all programs. And not just those which developers were not too lazy to recompile. WH> 6) possibility to create pair the processor and  the optimizer under it. What gain of productivity gives it there is no saying. Can be and in times. Points 1,3 and 4 it is enough , and hinders nothing them to embody life now at a stage of the assembly of a software (in clang), therefore they are irrelevant. Remaining it makes sense, only you exaggerate a role of the new processor and optimization under the processor in finite user-experience the user who has crucial importance. And meanwhile all it would demand still standardization of the intermediate code. Besides, to analyze sent by the user core-dump it would be more difficult. It turns out if to look theoretically at this question it looks beautifully. In practice of a pattern of a software and processors are arranged so that at a significant amount of efforts of favor would be a little: the computer in times would not be loaded faster, not  there would be applications, not pages in the browser faster would be loaded, and games would optimize all the same manually. I can present a noticeable scoring only in applications of handling of video. Here in 1990 and 2000 when every year frequency of the processor increased by 50 %, it would look it is useful. lpd>> I, of course, understand that the monopolism complicates technical progress, however it would be desirable to look at least at a prototype of the new super-processor before to change an infrastructure of installation of applications on billion computers. WH> we here consider a situation "all software disappeared". It is necessary to write from a blank leaf. Basically, the train did not leave, and it is possible to spread setting/compilation system and now - a question in demand. Want - do Linux with support of any processors and optimization at installation. Only the modern OS have lacks and is more serious.

295

Re: And if all from the beginning?

Hello, lpd, you wrote: WH>> Also is not present.  it is not enough. It generally is a microscopic part of all software. lpd> I hear words burned . I will not overpersuade, since it is normally useless. Simply on that that I know how many a software becomes outside of  a world. For example, that I now do not . And it never will be . Simply on that that it is not necessary for the customer. And on orders it is more than such software than . It it is simple for  it is invisible. lpd> points 1,3 and 4 it is enough , and hinders nothing them to embody life now at a stage of the assembly of a software (in clang), therefore they are irrelevant. clang it With ++. And With ++ goes wood from a word absolutely. lpd> Remaining it makes sense, only you exaggerate a role of the new processor and optimization under the processor in finite user-experience the user who has crucial importance. Users different happen. lpd> And meanwhile all it would demand still standardization of the intermediate code. You so speak, as though it something difficult. lpd> Besides, to analyze sent by the user core-dump it would be more difficult. On the contrary easier. For the broken storage will not be. Races will not be. There will be absolutely predicted end of process. After which it is possible to save  the process data in  a format. Pay attention  of the data, instead of crude storage. You understand a simple thing: the Architecture of the processor generally will not gleam anywhere. Developers of the processor will adjoin to it only. Well and those it is simply interesting to whom to play with it. lpd> It turns out if to look theoretically at this question it looks beautifully. In practice of a pattern of a software and processors are arranged so that at a significant amount of efforts of favor would be a little: Oh. Efforts in this case it is necessary much less than now. And generally all. lpd> Want - do Linux with support of any processors and optimization at installation. Only the modern OS have lacks and is more serious. Aha. All of them on With are written. And it is necessary to do something like . And to write in verifiable language. A problem at a present situation that under this system there is no software. Without a software this system is necessary to nobody. And under not the necessary system nobody will write a software.... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

296

Re: And if all from the beginning?