Sunday, January 24, 2010

Proshow Goldbackground

A community without private property

is not easy to shake off the image of "feudal user" that guy that shows credentials the doors of the wall, and once in their domains, have them as master. However ... conceive it otherwise.

USER SESSION does not directly involve

For a system which is accessible to several people at once arises the need for the concept of "session."

A SESSION represents a "working day" from the perspective of who is using the system via Terminal. At one point, N people will be connected to terminal N and therefore N must be separate sessions, one for each. If, in particular, all are using the same program, then there will be N instances of the same program which implies N separate data areas: one for each instance, ie, one for each session, or what is equal, one for each visitor.

Note that I said "visitors", not "user", this is because users need to define the concept derives directly SESSION but the concept of "private property" (to those users of the system on certain media, for example files). From the perspective of the OS, users are entities registered in advance in data structures designed for that purpose. In the presence of "private property", each session will belong to a user (previously registered) which will be identified (say who) before they can gain access to their own resources.

Another possibility is to give up the "private property." In such cases, all system resources are the heritage of all who visit and then the concept of user automatically loses its meaning.

SESSION concept, however, prevalce, it is still necessary that several people work with the system at the same time without interfering with each other actions.

Imagine all the people sharing all the World

The experience of a shared system without privileges to all visitors with us today on the Internet (Wikipedia is a good example), but try to imagine a system like Heritage / 1 in 1972.

Say a minicomputer Heritage / 1 is installed in a software production company. There are eight programs each with its corresponding terminal. All resources of the machine (especially the peripheral tapes, impersoras etc) are shared so that arbitrage should exercise any OS to prevent, for example, a print out with two servings intertwined different programmers.

The arbitration, however, does not imply property relations on resources but only a mechanism to bring order into licentiousness. For example, if two processes have requested the use of the printer, then one will have to wait until the other end.

Short of Anarchy

coloquemas Now a Heritage / 1 in a public library. The Library Catalog recycle in a database accessible to visitors through four terminals installed in the lounge. The visitor does not have to be identified, simply consult the catalog directly from either terminal, and not only performs read operations but also leave comments, suggest books not in the catalog, etc., provided on an "anonymous" (from system perspective) which does not prevent someone leave your name and phone number voluntarily.

Updating the Catalog, however, can not left to the visitors as this is a professional task, so that the system offers some programs for visitors and others to librarians.

This can be solved based on the terminal (port), ie without need for early adopters. The OS can create a table of available software for each of which specifies the terminals (ports) licensed to serve as an interface. In this way, visitors will never gain access to the catalog maintenance program.

How is this related to SESSION?

When the visitor enters the system through a terminal, you create a session tied to that terminal, the shell presents a list programs available to "him" (actually, for that port.) Apart from the catalog may have other utility programs such as math, the user could run these programs, get results, send to the printer (shared with others, you may have to queue) and then close the session or not.

terminals used by librarians operate similarly, except that they will have access to both the catalog and to the maintenance of the programs.

And what is the advantage of all this?

The advantage is not having to manage users within the OS, which represents a tremendous savings in terms of development effort and maintenance system.

We have seen that a system can be accommodated in a cooperative work such as software development and a public service with a library. In general you will find a place in a facility where the computer provides specific services to a specific audience, as was common until the early 80s.

Today, this type of service (purpose, visitors are not privileged) is still common, especially on the Web. Only that level of implementation itself uses the concept of a user for security reasons and, above all, because operating systems are used as implicit.

CONCLUSIONS

In a time-sharing system SESSION need the concept of building a "framework" for each terminal separately. The concept of users, however, is optional.

If not required "private property" (resource users), the concept of user is more. Implement only consiguirĂ­a imnecesariamente overload the system complexity.

If privacy is required the Administrator for certain system resources (eg, maintenance programs and configuration files), it is necessary to see the manager as a "user": just Restrict these resources to those terminals intended for administrators.

Does 1 Cigarette Harm My Baby

An account of experiences

While Heritage / 1 only exists on paper and half (although I'm building the first circuits), design counts as a personal experience which I have drawn some lessons.

CONTROL PANEL

could not imagine a minicomputer-70s without a control panel full of switches and lamps majable bit level. However, the design of this panel and, above all, your connection to the rest of the circuit have been presented so many difficulties that I even think of deleting it altogether.

In fact, my panel is much simpler than other "home-brew" minis that have been studied (See links), but even so it is recharged when compared with real minicomputers of the 70s as the DEC PDP-11. The difficulty of the panel is presented on two fronts: (1) requires a large number of connections to the CPU, and (2) can not be "smart" because it is the first place to look for failures to be so absolutely reliable.

the problems have been solved little by little at the expense of increasingly simpler. Finally I was able to reach a final design with which I am satisfied, not so with the large amount of time I had to spend.

As is teaching me a control panel should be as small and simple as possible with just essential for operating the machine, maybe not so much to help the development of programs. I did not want my current design to further simplify it, definitely, I have the experience to operate with all its switches and lamps. I learn from the practice of their resources which are useful and which not.



A general purpose machine

I set out to design a simple machine designed to address any type of work ... which is the same as saying no concrete work. In my article sixteen versus eight bits explored the idea of \u200b\u200ban optimized design as an alternative to general design that Heritage / 1 defintivamente took, for example, Heritage / 1 does not optimize the treatment of characters or flange support for a preemptive multitasking operating system with process protection.

The moral is that "general purpose" really means "large number of specific purposes." If I ever meet this challenge, I will apply this teaching to guide the design of certain "somethings", my effort will be led then to optimize for each of them.

HARDWARE ARCHITECTURE VS. SYSTEM SOFTWARE

One purpose of this project has been to provide an environment "understandable" where you can play at my leisure to design (or adoption) of system software. Although the architecture of computers has resulted apacionante for me, much more is being operating systems programming ... only that this can not penetrate until the hardware is not ready.

An alternative would be to design a feasible hardware to be assembled in a short time, for example using a microprocessor of the 80s generation as Intel 8085 or Zilog Z80, the cultural aspect had been satisfied with just "run the historical context" towards the decade. Only then I would have lost the "taste" of a machine with no mysteries, no "boot up" and without LSI, all that ultimately form part of the personality and culture Heritage / 1 .

DEFINITELY A "MAIN-FRAME"

The 80s brought the concept of "personal computer": a machine which justifies its use inexpensive dedicated to a single person. Today the computer has not only left the Computer Centre to the bureau, but this the pocket of the owner! "Indeed, a cell phone today is more powerful than a mainframe in the mid-60s as the IBM System/360.

current perceptions (both domestic and industrial) is that a computer must have a video card, a graphical interface and a mouse. Heritage / 1 vision carries the legendary anti definitely opposite: a computer only "computing" she be accessed through "terminal" outside, she, in itself, provides no way to be "look" or "commanded" directly by users.

A modern version of Heritage / 1 replace your serial ports (where the terminals are connected) through an Ethernet connection, and UNIX would show up as a "host" within a TCP / IP. The "terminal" would then desktops and Macs or PCs, or perhaps be manufactured proprietary terminals capable of presenting a graphical user interface, perhaps Web.

But always be, Heritage / 1 a centralized machine: a mainframe; not provide any means allowed to treat her like an ordinary desktop computer.

Thus it would seem rather to an IBM AS/400. My teaching in this regard is that the world of PCs do not find it attractive at all.

Saturday, January 9, 2010

Smothering Satine Socks

The real motivation of the project

Heritage / 1 is a project whose need is not easily explained in terms categorical and convincing (in that seems to smoking habit), but the effort to design, build and program a prehistoric computer, designed to perform well tangibly work "useful", demands an explanation!

Indeed, explanations have given as that of "taste past problems" and "honoring a legacy technology," . But if you touch the story with the hands they were would not it be more appropriate to clone a dinosaur of those, as did John Pultorak to replicate the Apollo Guidance Computer in his basement ?

At this point I must confess that my real motivation (underlying), without ceasing to be a technical and cultural nature, is more in my desire to produce a computer that can be understood, contrary to this "black box" on which I write these notes.

remember my first "emotional contact" with the program was in 1982, when he discovered eight-bit microprocessors, including Intel 8085. And my greatest joy that came from just that one could understand! 8085-A microprocessor no longer a "black box", but within the level of abstraction that provides the block diagram, one can work entrusted as in a boat sailing in shallow water. Not so with a PC.

Create a computer "meant" was because the main objective of Heritage / 1, hence the policy of simplicity in design. His ability to "useful work" was another premise, apparently in contradiction with the first hardware as "simple" will hardly bring fruit if is not at the expense of a very complex software.

The truth is that in six months, and even with a half-finished design, I have come to learn with depth on a myriad of topics on which he was just an idea vaguĂ­sima. Has been how to get to unknown lands, having given me the task of drawing their maps and, through them, eventually experiencing a tremendous familiarity with the landscape.

In that sense, Heritage / 1 has been, is being "a unique experience. And I think in the apparent vagueness of purpose, lies, paradoxically, their main asset: the culture, the component that does not support any mathematical formulation.

Technology and mathematics, will be cold and boring until acquiring cultural traits, traditions, ethics, aesthetics, everything that allows the designer intimate with your design in human terms, that is, with passion and joy. And so as a folder of drawings, a prototype and an idea -Heritage/1- have come to form a warm body who often live chat ... intimately, and it is this need for "privacy" that it requires simplicity and understandability. A unique experience, "I said only personally but as the technology is all full of traditions, no consensus on what a" design elegant "and what, a" pidgin, even when there are no formulas for the elegance and the macaroni.

Heritage / 1 has its own culture as: a story to tell, a tradition that follow, a criterion in the mirror. You can get to work or not, do it right or wrong, but in either case the criterion for evaluating its success is not in MIPS (million instructions per second) but their contribution to technological tradition which has sought join as an echo from a time when computers are fast, but not romantic.

Friday, January 1, 2010

Veyron Engine Drawing

Personality

The essence of Heritage / 1 is in its personality. Paperback, restrained , and simplistic, industrious and eager for challenges, but unwilling to change its anatomy and its way of dealing with life. The historical context (1966-1972) to which I have restricted it is, perhaps, just a pretext to avoid contamination with the sophism that essence exaggeration of our time. Heritage / 1 has a personality that justifies their name and their sense of self, and it's just that what makes a project which leads to intimacy. Personality

also means culture and Heritage / 1 involves the culture of the 70s who is also the culture of UNIX. My dream for her is give her a Minix operating system like Linux or BSD, and although I'm not sure that that happens, the spirit UNIX living in it from this nascent stage design that is still.

acknowledge, however, I've been a bit exaggerated to hardiness. For example, I may not use ROM even though the crest of UNIX time (1975 onwards) and "minis" used semiconductor memory and up to 8-bit microprocessors in providing support, for example, in video -terminal. If Heritage / 1 had really existed in the 70s, probably had been perceived as a machine "too traditional" and unadventurous design.

However, the roughness (simplicity) is still the main attraction, give it up would be to destroy it before you have made birth. My challenge, in addition to really work, which is itself a major challenge, is able to develop useful software capable of running on a simple architecture.

Scholars (or curious, like me) on these issues, you know that Bell Labs UNIX arose in the hands of Dennis M. Ritchie and Ken Thomson, two hairy at that time wrote their programs in teletype terminals using assembly language, like everyone else. UNIX was perhaps, according to a witness's own Dennis Ritchie, the first "time-sharing operating system" in history, or at least the first to achieve mass acceptance. And this is important because it is derives this spirit of collaboration that is part of its essence.

Indeed, before UNIX, programmers often write their programs to tip pen and paper. Computers working in batch mode: both data and programs came from punch card readers or magnetic tape and programmers were happy to send their listings to the computer center, wait for the operators loaded, run and proven, They then receive a list of faults detected.

UNIX, being a time-sharing system, allowing developers to write directly on the machine (as we do today), and interact with their colleagues, encouraging and cooperative work environment. In fact, UNIX was initially conceived as a development environment not as an operating system! UNIX

surprised, then, to their authors. The collaboration came out of Bell Labs to universities, especially California, Berkeley, where he made notable contributions. Long before the term "Open Source" was popular (or even exist), of Berkeley went tape marked "Software distrubution Berkely (BSD), a version of UNIX that continues to develop under the same name.

UNIX is a traditional technique based on respect for its past is a living example in "Man pages": even in the most recent versions can be read exactly to the original texts of 1972, the command "dc" (desktop calculator), one of the first utilities written for UNIX at Bell Labs

I can not ensure that Heritage / 1 ever run UNIX. What I guarantee is their loyalty to the tradition that his name represents. Heritage times / 1 are the times genuine Rolling Stone, "Pease and Love" and "KISS" (Keep It Simple, Stupid).