Good Server System

Edward Guo discussed his different opinions with me in my "drugs". I think it is necessary to add them to my favorites:

Edguo published on 21:39:46
Sorry for a few different points:

1. Linux Server creation is okay (FreeBSD was a good choice for the company's email server), while desktop applications are still poor ~ Five generations, redhat7 used to play a game, but the command line is nothing, but because the wingui virus is too heavy, I think the KDE/gnome thinking is very strange, I really can't stand it, it may be because you are too old to turn around. However, since the client needs to be selected (whether you are willing or forced), it is better to take a routine, otherwise the cost is high.

2. I think Windows Server is stable enough for most of our customers. It is easier to hire a Windows administrator, except for many patches. I think that the separation between the GUI and the kernel is more stable, just like the previous saying that 1 k of memory needs to be saved. The GUI kernel will become more and more stable, and the non-Gui kernel will become more and more unstable. this is my personal opinion.

3. Be practical and never be too programmer.

Raptor published on 13:55:28
1. I didn't say that Linux is better than Windows in terms of desktop, but it is much better than Linux in the past.

2. Of course, there is no problem with the internal applications of companies that do not require high requirements, but what about putting them on the internet? What about applications with large loads? The advantage of Windows TCO lies in the fact that Windows administrators are easy to find, but it is much less technically-system security and performance optimization work is not done by patching, the cost of a technically sound Windows administrator is not low.

Let us prove the relationship between the GUI and the kernel. This is basically the same as 1 K memory. In my opinion, on the contrary, the GUI kernel will only become increasingly unstable. Otherwise, there is no need to set up a protection mode for the CPU-this will undoubtedly add additional overhead.

3. What I am talking about here is intended for programmers. users need to know it by themselves. It is wrong. They also find it by themselves, but programmers should know how to be nice to themselves. ^ O ^

Edguo was published on 16:47:12 IP: 58.251.82 .*
1. Another sentence: cost. The technical routes of servers and clients are different, causing high costs.

2. I know that China Merchants Bank's online banking has adopted a large number of Windows servers. You can attack it (but maybe I am out of it and they have already changed ), our customers (a provincial-level department) also use Windows server over the Internet. For large loads, please test it on Microsoft.com (the impact of today's earthquake is not counted, haha ).
Windows Management really does not need a high level of quality, of course, you can take a street goods as the administrator is not good, I really don't think the requirements for the configured Windows Server are high (maybe my own level is low ). In the past, I saw people in the opposite group engaged in IBM hosts and I felt that their maintenance costs were too high. Although I hate M $, I still need to admire it.

I can't prove whether the separation of GUI and inner core is getting unstable, but windows is getting more and more stable and everyone can see it (that is the improvement of Design). It is a fact that I don't think it is a coincidence. 1 k of memory was used by everyone in that year. In, billgates also said that K of memory was sufficient for anyone, but with the huge expansion of memory, it makes no sense to save that 1 K. If I think I keep up with M $, if it is not necessary, I will not consider the so-called stability to switch to other platforms (unless it is enough to attract customers for example, I can make more money using Linux, and the customers think it is very worthwhile or the customers want me to use Linux, I don't have method ). I think most of the so-called stability depends on your software design level, not the operating system or compiler.

The protection mode has nothing to do with the GUI. If yes, Linux does not have to use the protection mode and runs some junk software without knowing whether it will be "stable ".

3. Too many programmers mainly read those posts and do not target the fierce brother. In fact, I certainly have nothing to say about these nonsense, and I don't have to think that I have been dedicated to Linux since I took the Windows route. For example, when our programmers review the functions, I asked him how to do the interface functions you provided now if the customer wants to do this, he replied that I want a-> B-> C->... -> K. That result is often given to K. We have been working as a programmer for so long (since the first time I made money by writing a program, I have been there for 12 years). People who have comments here may also feel better technically (including me ), but it is really a good technology to eat, and many programmers (even myself) are always enjoying the technology, if you do not put the interests of customers, projects, and products at the top of the list, you will suffer losses. Everyone knows that programmers are still willing to continue to eat this kind of loss when they eat young meals?

Wang haihan, the author of the offense.

--
Edward Guo

Raptor published on 14:57:23 IP: 203.94.2 .*
1. I don't know what you mean, what is the cost? The cost of the server is inherently higher than that of the client.

2. Since they are all banks, I certainly don't mind using professionals. I also admit that Windows is well configured and the security is equally good. There is nothing to say about Microsoft websites. Windows is developed by them. Of course they can be used better than others, but not all users have the same strength-a better example: The NASDAQ website also uses Windows servers. However, this is meaningless. Just make a brief summary of the large portal websites. This is a typical big load application. How many of them use Windows?
You don't think the requirement is high because you have reached a certain level, but in fact it will not be too difficult to maintain * nix servers after the technical level is reached, just a habit process.
Windows is getting more and more stable, because the previous windows is too unstable, but there is still a gap with * nix. People with even higher levels will inevitably make mistakes. Once an error occurs, the risks in windows will be significantly higher. In contrast, * nix provides more comprehensive protection. Gui and kernel isolation are only one aspect. In this aspect, the GUI errors of the next user process in * nix are only displayed as a Socket Client, however, windows may cause serious errors in the GDI part. The memory size problem is not the same as this one. So I also introduced the protection mode as an example. The purpose of the protection mode is to isolate the OS kernel and user processes, when a process error occurs, the protection interruption occurs, so that the system kernel can control the status in time and improve the system stability. The separation of GUI and kernel is essentially an extension of this idea, while windows closely integrates the two in order to display performance (desktop applications have a strong demand for this, this will bring some obvious risks.

3. I have not been using programs for a long time, and I have not been a full-time programmer for several years. Now, my identity is equivalent to the customer's system maintenance personnel. However, if any software vendor makes the software that your team can implement, then I also need K of them. However, if you think that * nix applications are as you said, it is your misunderstanding. In many windows, you only need a line of command under * nix to perform n mouse operations. The only difference is that the operation method is different.
The good-looking interface and intuitive operations are certainly an aspect of the customer's interests, but this is more superficial. When a system is running for a period of time, when the customer starts to perform maintenance operations such as restarting from time to time, the customer's benefits are different.
A real server is not a simple system to maintain, but a system that does not need to be maintained in most cases.

Edguo was published on 10:40:51 IP: 58.251.85 .*
This refers to the removal of design and other factors. From the perspective of development and maintenance, you may have four employees, one on the client to develop Windows programs or pages, another person needs to develop non-Windows programs (such as Linux) on the server, and the third is the person who maintains the server, the fourth step is to maintain the client. (do not tell me whether B/S is zero-maintenance. Most of the maintenance tasks are not the client itself, not only does it involve cost management (such as staff replacement), but it also involves employee development and other issues. If you can explicitly reduce the cost of one or more of these aspects, it is the master.

Sorry, I made another mistake ~~~ In fact, I am not that old, because the first time I made program money, I went to college ~~ I should belong to the kind of programmer transformation that is not well done, and now I have to work as a programmer. So I strongly advise my uncle at the university to take a major role in computer science, and to take management, economics, humanities, and other social sciences, that is the weakness of programmers.

In this regard, Ling Hu also sent his opinion to me:

I have never worked as an administrator, but I have been using Linux for a while, while LP has previously worked on Windows system management. Therefore, I would like to say a few words from the perspective of use.
A good Windows System Administrator is never a simple task. When you manage thousands of computers in a company, writing automated tasks is indispensable. At this time, Linux is much less difficult than Windows, because on Linux, you still face commands you are familiar with, just use shell to combine them, even remote execution only requires ssh. In Windows, you have to get new access to a tool called "host script", which is different from Console Commands and window operations. This is an additional knowledge that requires additional learning at a cost. -- My LP friend is in another company, whose System Administrator does not know this. It is a manual operation and a virus attack occurs in the company, my LP friends stayed up for N consecutive days. (This is actually happening)
What's worse, host script uses com. If there are no ready-to-use components in the system and you need to develop them on your own, it is more difficult than developing a command line tool in Linux, and the probability of errors is much higher (because of the complexity of the COM model) -- This is usually not done by a system administrator, so it is a cost to ask a com developer.
In short, my biggest experience in Linux is his kiss principle. Windows has made maintenance and development too open, either silly operations or API programming. In Linux, with the powerful adhesion capability of the script, it provides a very good intermediate level. At this level, you do not need to be very proficient in development, but it can be much more complicated than "silly. This will bring great convenience for management and maintenance.
In addition, in terms of security, Linux may not be more secure than a well-configured Windows, but the benefit of Linux is that it will let you know the security from the beginning. If you don't know what user permissions, user groups, and owners are, you can't do a lot of things, so you have to understand them from the beginning. The most terrible thing about Windows is to hide these complex concepts behind ease of use-even if you do not know anything about them, it will not affect your use of the system. Therefore, many people have no idea about security, but think they are familiar with windows. This is the source of Windows security problems. However, it is not easier to fully understand windows concepts than to be familiar with Linux.
In addition, I just did not say anything.
In my opinion, Linux is not more difficult to use than Windows as a desktop, or even more convenient than Windows in some aspects (such as connecting to my a1200 mobile phone ). The key is that you cannot use Linux using Windows. In Windows, if I want to do one thing, and this is just because Windows cannot do it, I will find a software on the Internet. Maybe it is not particularly suitable for my needs, then, I can't help it; I can either open the development tool and write it from the beginning, but it is generally difficult and has a long cycle. In Linux, I will find a program that can complete core functions (such as basic functions such as graph and editing), and then write a script to combine them. After the script is written, it is easier to use than in windows. Writing scripts is much less difficult than programming in windows.

For the moment.