Open Letters #01
20 июня 1999 |
|
Reply to the letter number 4.
08/27/1998 Surgut Hello, dear Captain Nemo! Wrote you a simple Spectrum coder. Since you are a circuitry Speccy, I would like to introduce you to my thoughts on the future platform SPECTRUM. I am sure that no one in this business and try to take this into account. 2. So, for today the priority ways of development are as follows: Hardware: 2.1. The development of pure processing power without losing compatibility. 2.2. The sharp increase in graphics capacity. Music and digital left after the GS. Software: 3.1. The new operating system with auto-tuning all the hardware specifications of your computer and peripherals and uses all available resources. 3.2. The emergence of powerful systems in the area music and graphics (MULTIMEDIA). And the general problem of standardization. 4. I agree with many projects (Splinter, GMX, Transformer), but with my point of view, the emergence of the "monster" will result in the following problems: 4.1. Lowering compatibility. 4.2. Neokupaemost project (the user can not simply buy another one computer, even if the cost of unprofitable for the seller). 5.Vot my ideas (a sort of hybrid of all what has been said on this subject). 5.1. The sharp increase in processor power without losing compatibility (this is rather a question for you - whether it is possible to make is a purely hardware?) 5.2. "Accelerator" is inserted instead processor through the plume, and the processor, who left their homes and inserted into "Accelerator". By default (after reset) is active Z-80. Any port switches the current CPU, RAM and bank Rom. in ROM can write the initialization Z-180, BIOS, AUTOCONF, etc. 5.3. Note: (*): registers can be transfer and programmatic way. 5.4. I unfortunately do not understand in electronics, but still think a little logic here is. 6. Schedule. Here, almost everything is in GS. 6.1. By default (after reset) is normally installed Spectrum screen and is determined by available video memory. In the ROM protection control program procedures for graphics operations of 4 types: - Text mode; - Bitmap (block); - Vector 2D (lines, circles, shading etc.); - Vector 3D (three-dimensional drawing objects, 3D scaling, etc.) 6.2. It would be nice if present: - Digestion BMP, JPG, TIFF, PCX and etc. - Unpack the packed data (Graphics). - Math functions. More important to make such maps straight memory access (in this case - video). Well, in RZU stored variables. In addition to the built-in, would not prevent support for arbitrary modes (dXxdYxC) and palette (within memory). Having the opportunity to (unlimited) memory expansion and replacement of the processor. 7. Here, perhaps, and all of the gland. I myself am now develop powerful (For Speccy) graphics editor. It will be implemented before unrealized Speccy at: - Vector FONT'y. + Blend ("Moss" techturoy) + 16-bit sequence "Shape" of the line (the dotted line and TA) + REAL TIME MAGNIFI X8 - Editors FONT'a, "shape" of the line texture points (8x8), etc. + 2 Screen - Connection instead of a point on any object (line, circle, box) + Error-free pouring - Scale, perspective, tilting, twisting, etc. "+" - Has already been done and working. Also already made all the graphics primitives. All operations are carried out with maximum Spectrum for speed. In total, about 30% done. The length of the mash. code Editor's total will be about 40 Kilobytes or about 12000 teams (it's the eye). Now somewhere in 12k code. Given the scope of work and what I one term out of more than unknown. Especially because for many operations have to develop new algorithms (or nowhere take, or have not arranged their characteristics). Thus was born new algorithm for drawing a line (the part procedure line, which will decide - to move or not move to point down represents the 11 (!) processor commands the Z-80, including the only transition team JP M - it runs as a result of reflection. The same algorithm is used when zooming slope, etc. I would like to know the opinion of g.Brezenhema about it. 7.2. Also been developed entirely new algorithm for the fill. It has five major advantages: 7.1. Clearly set the buffer size (2 * Max size of the fill for Speccy 2 * 6144 = 12k). 7.2. Great speed (on the Speccy more than 30000 points / second). 7.3. Correlation between the complexity of the terrain and time casting tends to zero. 7.4. Ratio is less than Max size of the area, the higher the rate of "pouring" is hermetically sealed, ie, if all screen (49,152 pixels) is filled with 68 frames (1.36 sec.), The window for the vector letters (64x64) will fill in within the next 5 frames. 7.5. Full infallibility. This algorithm eliminates errors by 100%. The implementation of this algorithm on a Speccy WARRANT fill any object within 5 (five) seconds. Although the rate can be increase somewhere in the 20-25%, but we need to convert the entire screen in the format line for Line and take another 6 kilobytes. All figures quoted for the fact that processor operates at a frequency of 3.5 Mhz. That's all that I could write. 8. Dear Captain Nemo! I would like to know your opinion on all of the above and at the same time get the latest price lists. 8.1. P.S. Can I patent algorimy if - yes, how? 8.2. P.P.S. If you want to review the above algorithms, I You present them with a complete description of their work. 8.3. P.P.P.S. I apologize for all the I admitted errors or misprints. Yours sincerely N. 11/09/1998 Surgut 26.11.98 C Nemo Hello, N! Apologies for the delay in response, because on letters of this kind is very difficult to answer. Wrong the approach to the problems Spectrum. According to formal logic, of incorrect attempts (False - Engl., damn - Lie, cheat, deceive stupidity - Old Russian) may follow anything, including real approval. However, everything in general - wrong, despite the fact that there are true statements. You can also see that you too are abusing viewing TV (television) and reading the mass of the "yellow" computer books. Your main mistake - it is confidence that the underlying cause is the growth of computers and their technical parameters. In your opinion, is cause, "prime mover" of all that happening in the computer world. This is not so. Specifications - this is consequence of the development platform as a socio-cultural (or cyber-social) entities. Is only as good structural development platform and how it is mass, what problems are solved on a platform - gos.upravlenie, professional and business activities (IBM); demomeyking, kulibinstvo, study computer equipment (Spectrum), - depends on technical "tricked out" equipment. Pointless to put the cart before the horse and deliver microscope that they hammer nails. "So, today the priority ways of development are. " 1. Changing spectrum of platforms Spectrum. As through the introduction of mass users of new applications Spectrum and new methods and fields of use, and by improving existing structures and areas of use. 2. Evolutionary, continuous development platform. With this approach, the pace of development is determined not "tricked" the coolest "monster" (and their naplodili many), but the intensity of the replacement of obsolete equipment. Transition to new, improved standards and the rejection of the old, exhausted themselves. From a financial point of view can and should do only what is in unable to pay the users. Ie on clothes to stretch legs. Senselessly invest in the poor, they never return. 3. Development ideology Spectrum. To every lamer knew that Spectrum - a TV and only TV as a display device. That this can not be changed will have trouble if everything is throughout the community. And here, for example, tape interface can and should be removed. That work should be "under the" operating system "and not" with it "or" over it. "What operating system must be one. (Imagine that in addition to TRDOS'a would exist and all the other disc interfaces, "naplodilis in Europe? Spectrum would simply not take place as a platform). Now let's try to analyze your deals with these positions, not based on from the conventional foolishness "Spectrum'u lacked the technical characteristics." Thus, by Settlements: 2.1. Let's develop a "pure processing power" (meaning, apparently, the clock frequency). Now machines competently made Turbo-mode units percent of the entire park. Programs under They do not - and therefore nobody knows what there is a turbo-mode is made correctly, and not through the ass. So write a program which would use these features. This will be the development. 2.2. For graphics power is required display (namely the display, not TV without channel domestic production, which is lamer called "display"). Try to sell at least a couple, and then free to serve them - money that users do not have! It is also useful for exacerbating the problem of understanding to sell a couple more screens in the mail in the village, say, Tmutarakanku, where the voltage jumps in the region of 150-300 volts, and see how they're technologically advanced. 3.1. Let's make a new operating system. At my expense, for your money. Pay something no one else. Spend 10-20 man-years to build, debug, and polishing the code, print (and most importantly, extend!) 50 tons of commercials accompanying literature, will buy shares commercials with the twenty- Spectrum clones (and they are just so much on expanses of Russia). Tenancy for their installation. Zaimeem phone, PO box (this is necessary to reverse communications and customer support - and how You'll learn where the system is buggy?). During smoking breaks in this seething activity we are on the windowsill in the office procedure of combining business with pleasure to write the first version of the operating system so that it value can be ignored. Probably it will cost thousands of commercials 300-500. Requisites indicated in the price list as soon as transfer money, so immediately and get started. 3.2. And what powerful systems of music and graphics (MULTIMEDIA - what a term! On TV must have heard?)? There Amig'a. There is IBM. Generally speaking, there are and GS. Problems of contemporary music rooted not in the imperfection of music instruments, and the imperfection of the listeners. Recorded precedents when not only the performers enjoyed soundtrack ("plywood"), but the auditorium. In addition, the "harder", "heaped" equipment, the more code (the program) to it. As in the cost (ie costs for its creator), and in the termination costs for buyer. Dilettante (amateur) approach to programming in this case is no longer possible. If you are in single-handedly trying to create a product then years of commercials are 15 you will take. In structures (firms) in Spectrum'e virtually none. 95% - is Kulibina as home, financed by parents. Even if we assume the impossible, that these 15 years you honestly worked - who will buy? Even are fairly simple software products marketed on a commercial basis, are paid what is called "through the stump-deck." But there are also pirates ... 4. Splinter and GMX - it is really projects. Transformer - a conceptual development. Ironically, the conceptual design may be more realistic than the draft. If the "Transformer" was project, it would undoubtedly be found "Monster". However, if you read the beginning pismp, we find that the Transformer can run, but Splinter and GMX can not even be in the future. 4.1. Lowering software compatibility is inevitable. Legal: 90% transitive software compatibility, 95% of direct software compatibility. In this development will be smooth, wrinkle-free and ambiguity, ie, continuous. Be a Transformer implemented now, the relevant parameters would be accounted for 60% and 50% respectively. Total have: 0.6 = 0.9n. Solve with respect to n: n = lu (0.6) / lu (0.9) n - 5. Ie requires approximately five intermediate models. Given that a number of reasons the model can not be replaced more frequently than one year and less than two, we get the commercials years 5-10. And most importantly, the market should be ready for the real application of this technique. So the main problem in the brain and not in glands. 4.2. What neokupaemo within a year, payback within 10 years. And the user does not necessarily buy another computer. This is absurd. According to clause 4.1. he will have to buy 5 computers! Necessary to change the board and make upgrade. in which case the user will be regularly funding the manufacturer. KAY series has already upgrade-chain of the 5 models (RAY-256/2SL-94 city / V 1.0 to KAY-1024/3SL-TURBO- 1998). The difference in price between the models is in the cards average of 100 rubles. In this case, all the boards work, and only the first model begin to be disposed of. This alternative "completions" (kitchen) approach. It is easy to see that after 3-completions (with approximately 50% probability) PC flies in the trash. Repair it impossible to sell as well. Here this is getting to the money, because need to buy a new PC (1000 p. - compare with 500 p. - By upgrad'u in installments for 5-10 years). Since users have no money, then they are no longer those of be. 5.1. Count This is all. And, most importantly, understand: only a quick cat fuck - everything else is going on slowly. As for figure 1, then it's something painfully familiar reminds me - I'm not GMX right? You take the floor barn - which is already about to fall apart - and make it two stories. Illogical. Under the longer term, apparently refers to the third floor and attic. That's altogether from the realm of fantasy. But it is possible, - In Russia we all can - she is always famous for fools and roads. The first is not fewer, while the latter did not become shorter. One thing to consider, this is how you explain the simple fact DEVICE that in Figure 1 will be more expensive similar, all-new motherboard? As reasonably possible chided the user on his barn and say: "Do not cry, even will serve you well for him rooted in heart, each mikroshemka you in it path. Let's also cooked soup from an ax. "Here and get the money. 6.2. Figure 2. Yeah, that's attic. Do you know how much it will cost this unit? Only one box, discreetly tucked away in a corner of the diagram, namely pryamougolnichek decoder PAL / RGB "will cost you one ruble in 300 commercials. In corporate Spectrum'ah TV-interface (ie, hardware bells and whistles, are responsible for connection to TV) takes from 25 to 30% of salaries (and, accordingly, the cost). By the way, but what comes out of the charts arrow to the BUS? Completely finished construction, however. Spectrum no longer needed. While there, I'm wrong, nneobhodim cause it to vparit. This is where the bus and useful. According to clause 6.1. Listen, why all this need (meaning a list of item 6.1)? Give an example of at least one real goal-setting, for which someone will be willing to give money. There are not talking imaginary, and ephemeral features (so to speak, enticed mental images), and real application? However, very satisfied with the familiar set of words from you computer industry. What will be the result at the level of the platform? That useful it will bring to users? That they will do after they get bored to solve this puzzle and try to somehow use a little? 6.2. About the digestion of BMP, JPG, TIFF, PCX, etc. skazt nothing I can not. Here you have surpassed me. These terms I not familiar with. Professional characterized not only because he knows that he needs to work, but does not know that it is not necessary. Loan to bother with this nonsense? all of clause 6.1. and 6.2. I note the following. Architecture are not additive. This means that "the best is the enemy of the good." If put it more academically, that architecture, like any complex system is not necessarily optimal, if compiled entirely of the optimal knots. And vice versa - architecture (structure) may be optimal, even when composed in part of suboptimal for the individual nodes. If collect all possible advantages of different architectures into one, then get one a major drawback. Pile - it is a heap. This, incidentally, most importantly, do not understand Spectrum developers. They gathered all the wishes to ONU lot lamer. Heap and turned. And finally. When a person require unlimited number of brain (Memory), it casts a dark shadow on the on his own brains. Computers, understood as architecture, complimentary mentality. This is a syllogism. 7.1. With respect to software - and here I am no expert, and not professional. Obraschayts in firms that engaged soft'om. The only thing I would like to note: item 7.3. should be reformulated: "Time to die is almost independent of the complexity of the relief." 8. Algorithms for a patent meaningless. The patent mechanism is not designed to guarantee pension for It happened once seen smart idea of someone's head. Copyright protects the particular implementation, not ideas. Between these concepts is a difference. Essentially, this is investment protection (Capitals), sent in high-risk areas. Under the capital (investment) can be understood and your personal time (Labor) spent on product creation (Finished product, not the ideas!), And direct funds, and overhead, cost of oboruzhovaniya to get the result and etc. etc. Ie main purpose of copyright law - is not to provide someone a comfortable old age, and investment protection with risk and contribute to the transformation these investments in the capital. Should not be confuse money with capital. Money - is just money. Capital of the same - it works in a certain direction the money - like working body. Tedious to give detailed legal interpretation to bringing unprecedented. Try to understand this in the form as stated. With reference to your letter should be noted that the finished product is unlikely Does not far off. Legal right to dispose implementation (product) has one who has suffered major costs. In particular, firms in the products of intellectual labor will become the property themselves firms, not those who are under an employment agreement, this product have acquired. As for the means of protecting copyright, the State is not up to this, and rescue of drowning - was the work of drowning. Here is an example of effective protection of their copyrights. When the market Kazakova (St. Petersburg) there were "left" version of the "UFO" Glory Mednonogova, it is not hesitation, went to g.Zonovu and told him about the following: "If the sale left-discs will not stop, then the next version at Scorp'e will not work. " g.Zonovu had driven away empty-handed. Indeed, bad shit where zhresh. In the apparatus used more complex combinations - they long to explain. The idea is that eventually hit falls on a pocket of who "Stolen."
Other articles:
Similar articles:
В этот день... 21 November