1

Topic: At once and well

For some reason this thesis that it is necessary to do all at once and well that it was not necessary to alter, obeys it is very tasty, and it would not be desirable to argue at all. But on the other hand for software development same the terrible thesis, I represent how many money it is necessary to make all at once and well since the concept well very quickly changes. For example 100 MB the distribution kit were more recently very bad and to make at once and well it was necessary to spend a considerable quantity of resources for its optimization. This thesis can went from any other areas, for example from building and developers of its steel to repeat. Whether you agree that it is necessary to do all at once and well that it was not necessary to alter? And in  projects it turned out to reach it?

2

Re: At once and well

Hello, okon, you wrote: O> For some reason this thesis that it is necessary to do all at once and well that it was not necessary to alter, obeys it is very tasty, and it would not be desirable to argue at all. But on the other hand for software development same the terrible thesis, I represent how many money it is necessary to make all at once and well since the concept well very quickly changes. For example 100 MB the distribution kit were more recently very bad and to make at once and well it was necessary to spend a considerable quantity of resources for its optimization. It so depends on the project that discussion is not meaningful. Whether O> you Agree that it is necessary to do all at once and well that it was not necessary to alter? A problem that frequently is not known how it - is good. Therefore the system should be floppy and admit painless refactoring.

3

Re: At once and well

F> it so depends on the project that discussion is not meaningful. And it is possible examples when it depends on the project. Speech certainly only about software projects.

4

Re: At once and well

Hello, okon, you wrote: F>> it so depends on the project that discussion is not meaningful. O> And it is possible examples when it depends on the project. Small projects with precisely fixed requirements can be made at once well. Very much helps, when the command does that already knows how to do. With large-scale projects of type of OS where requirements float or are not defined till now, unreal to make all beautifully.

5

Re: At once and well

Hello, neFormal, you wrote: F> Hello, okon, you wrote: F>>> it so depends on the project that discussion is not meaningful. O>> And it is possible examples when it depends on the project. F> small projects with precisely fixed requirements can be made at once well. Probably, but in my practice something concerns such fixed projects from the decision of any specific application-oriented task for example report formation, or any robot on data array handling, i.e. that that can be made for 1 , and that happens that it is necessary to add something there. As a rule such small aaplets are written on the contrary badly with .. Design, without tests, classes and layers, but quickly. All that is more, for example the interface from 5-10 forms, is as a rule already strongly subject to changes in due course both it is difficult to write at once and well F> very much helps, when the command does that already knows how to do. F> with large-scale projects of type of OS where requirements float or are not defined till now, unreal to make all beautifully.

6

Re: At once and well

Hello, okon, you wrote: F>> it so depends on the project that discussion is not meaningful. O> And it is possible examples when it depends on the project. Speech certainly only about software projects. Well here to you two extreme examples: 1. Start of the companion or a space probe. Requirements to a software are known to make in advance all well - it is possible. Thus start will be one, in it the heap of money is enclosed, and possibility of modification of a software after start will be impossible or is very strongly restricted. Here it is necessary to do at once well. 2. Business wants to make to itself operation automation, but at them the certain chaos, processes not rigid though at the same time want them to formalize that is why requirements to system will be specified even after its implementation reigns. As a result here the system will be unambiguous to be specified and altered during the life. Well and in between there are any average cases.

7

Re: At once and well

Hello, fmiracle, you wrote: F> Hello, okon, you wrote: F>>> it so depends on the project that discussion is not meaningful. O>> And it is possible examples when it depends on the project. Speech certainly only about software projects. F> well here to you two extreme examples: F> 1. Start of the companion or a space probe. Requirements to a software are known to make in advance all well - it is possible. Thus start will be one, in it the heap of money is enclosed, and possibility of modification of a software after start will be impossible or is very strongly restricted. Here it is necessary to do at once well. Here only it is not clear that means well, for example in case of 2 it is good when easily to make change when the code can be accompanied. But in 1 the companion these qualities are not so necessary, it is good in this case when the code simply works, can is not so optimal, is not scaled, but it is enough to fulfill functions. I.e. By and large debugged  in one function here quite approaches. F> 2. Business wants to make to itself operation automation, but at them the certain chaos, processes not rigid though at the same time want them to formalize that is why requirements to system will be specified even after its implementation reigns. As a result here the system will be unambiguous to be specified and altered during the life. F> well and in between there are any average cases.

8

Re: At once and well

Hello, okon, you wrote: O> Here only it is not clear that well What difference means? You ask not what such is good"and what to do" is good at once "? Well so it is not important, as it is defined well for a case in point. Important - you can make then changes to correct that initially turned out"not so good"or not. Or, generally, will be how much problematic to make these changes then.

9

Re: At once and well

Hello, okon, you wrote: O> it is possible, but in my practice something concerns such fixed projects from the decision of any specific application-oriented task for example report formation, or any robot on data array handling, i.e. that that can be made for 1 , and that happens that it is necessary to add something there. As a rule such small aaplets are written on the contrary badly with .. Design, without tests, classes and layers, but quickly. It is bad to be you. And it is is ineffective.

10

Re: At once and well

Hello, neFormal, you wrote: F> Hello, okon, you wrote: O>> it is possible, but in my practice something concerns such fixed projects from the decision of any specific application-oriented task for example report formation, or any robot on data array handling, i.e. that that can be made for 1 , and that happens that it is necessary to add something there. As a rule such small aaplets are written on the contrary badly with .. Design, without tests, classes and layers, but quickly. F> it is bad to be you. And it is is ineffective. Why, if the customer is happy with periods and operation Or it is bad that that the prices I force down

11

Re: At once and well

Hello, okon, you wrote: whether O> you Agree that it is necessary to do all at once and well that it was not necessary to alter? "The gram, Degree, Copeck" on my storage still was possible to bypass to nobody the law of the Sanitary technician. For us it sounds as "Select two of three: time, quality, resources.". Since in a subject it is a question of quality, it is obvious that to make at once and it is well possible only or at unlimited resources, or at unlimited time. And that, only in perfect conditions when, for example, banal nonsense of managers, architects and developers aspires to zero.

12

Re: At once and well

Hello, okon, you wrote: whether O> you Agree that it is necessary to do all at once and well that it was not necessary to alter? Here small , . That also everything, and at once, moreover and is good - such does not happen, sm on triangle MSF in presentation IT the Author: IT Date: 15.02 01:59. As a rule, when speak about "at once and well", speech normally goes about the politician of control of a technical debt the Author: Cadet Date: 01.06.13. Or we do not admit its origin, or we build development process so that it was not significant risk, or as a result we stick in "to make A, it is necessary to repair at first B but then the C" falls off. All secret of the master consists in handing over the project before the third variant

13

Re: At once and well

Whether O> you Agree that it is necessary to do all at once and well that it was not necessary to alter? O> and in  projects it turned out to reach it? "Quickly, qualitatively, cheaply. Select any two." () I would tell so. Key is a development of new system or finishing of the existing project. If the new project it is possible to make at once "well". Though often happens that those who come for you with it disagree. If finishing - that is superimposed at once a heap of restrictions. And normally anybody does not allow to rewrite completely operating system. Accordingly at finishing at me a position 1. It is necessary to represent - as it should be and as it should work. 2. Any change should approach to an ideal pattern of system. 3. Exceptions of point 2 are possible - when it is necessary to make something one-time and quickly. But when come with this one-time in the second time (and such happens in 90 % of cases), should be made on point 2

14

Re: At once and well

Hello, Sammo, you wrote: S> 1. It is necessary to represent - as it should be and as it should work. S> 2. Any change should approach to an ideal pattern of system. S> 3. Exceptions of point 2 are possible - when it is necessary to make something one-time and quickly. But when come with this one-time in the second time (and such happens in 90 % of cases), should be made on point 2 Simply here often confuse "should" from the technical point of view and "should" with ..  For example any operation which becomes 100 times a day is fulfilled for 10 seconds. I.e. somewhere in general runs for day of 15 minutes of working hours. If to look at it from the technical point of view at that it is possible to see that there is admissible the effective algorithm is used not and the correct fast algorithm "should" be used therefore operation will be fulfilled <1 seconds. But with .. Business it can be and it is not necessary absolutely if 10 seconds on operation do not hinder business, for example at you a flow of clients such that clients address time 5 minutes and 10 seconds here do not give capacity reduction. And with .. Business the ideal pattern of the world is to make not expensive and sufficient algorithm not is mandatory correct from the technical point of view

15

Re: At once and well

Hello, okon, you wrote: whether O> you Agree that it is necessary to do all at once and well that it was not necessary to alter? It depends on the project. If the project idle time, typical and in it it is from the very beginning clear as what to do more cheaply and it can faster appear to make at once all correctly. If uncertainty level is great, attempt all to plan and make in advance at once in better type with a high probability the project finishes off. The management skill the project partly in that also consists, that for the specific project to select balance between  and adaptability of a control model.