Shulou(Shulou.com)06/03 Report--

This week, the child was sent back to my grandparents' house. I finally have time to update my Weibo. I haven't updated it for three years. I am really TMD lazy! I was wrong. Try to update something this time.

Some people asked me why I didn't go to some of the bigger Weibo sites to write about it. I don't think it's necessary, anyway, it will be crawled by search engines sooner or later, it's the same everywhere.

This article is a pure theory, is an ideological guidance, you completely copy, you lose. As much as possible to write some of the elegant and popular, together with the study of learning progress!

The text begins. What a lot of nonsense. Sure enough, one is getting old.)

First of all, we need to clarify two concepts.

Waterfall development: waterfall, as the name implies, from top to bottom, continuous, steady progress. Waterfall development is one of our most conventional development methods, which is used in a variety of industries. Before we do it, let's arrange a meticulous work plan, which is interlinked, with pre-conditions, and the work to start on the most day. Then move forward steadily, and finally see the results.

Agile development: this is a development model that has been popular in the first five years. There are many genres. Agile is Scrum here. Agile development is a bit like a person who travels around the world without a map. You are moving forward and changing your route all the time. Of course, this metaphor is a little inappropriate.

I still don't understand. Please refer to various wikis.

OK, let's go back to the online game project, which is also a kind of responsible IT project. Although his participation functions are super many and the cycle is not short, it is natural to apply some development guidelines.

First of all, we have to assume a few premises (why? Specific analysis of specific problems)

1. Our project is brand-new and without any foundation (it can be a new platform or a brand-new creation of game types, anyway, there is no accumulation, or there is accumulation, but we can only look at it)

two。 Our project needs to be optimized, skinned, upgraded, filmed, and platform transplanted (anyway, it is not a new creation, inheriting existing facts and making the same type of products)

3. I'm going to use our RPG 3D engine to play games or something that has nothing to do with RPG (borrow existing resources or third-party resources, do something you don't accumulate).

These are two obvious problems encountered at the beginning of the project. OK, let's analyze them in detail.

1. First of all, if you are a brand new project and are faced with the awkward situation of being poor, why are you agile? Agile emphasizes continuous delivery and continuous embrace of change, and the biggest problem with Agile is that there is no long-term goal planning, which is not suitable for this 1-2-year long-term development period.

In the poor stage, waterfall development is your best choice, you should strive to build your product concept prototype, or steadily build your basic engine (server + client)

In the waterfall model, you have a very clear goal and a clear phase and milestone. There are generally no directional errors in these basic contents (of course, if your planning is bullshit and what you want to do is wrong, then SB, no development model can solve this problem. Of course, there is one thing that can help you. Please refer to my follow-up blog about the establishment of a standard online game development model)

Well, as long as we carry out the pre-planning of the project according to our development model, we can get a highly enforceable work plan, whose goal is very clear and clear, and there are a large number of experiments in it. and to achieve a very complex architectural design, anyway, generally speaking, we have done a lot, that is, we are doing the foundation, and there is no shadow in the game at all.

Well, after half a year of efforts, our basic platform has finally been built (server and client engine, non-business logic part, non-business support engine), and we are about to start our product prototype. We need to use our semi-finished tools to develop prototypes and try to build our flower shelves according to code-driven. If the result is built and the boss is satisfied, then it will enter the actual development stage.

The actual development phase is not used in the previous stage, it is real, code-driven, process-oriented approach, no, or you will die later! Now need to turn to data-driven, the whole re-hit a real shelf, after this stage, I think it is time to consider changing the way of development!

Why? Because once the real development prototype of the game is built and the core system development is completed, you will be faced with a serious question: is your game fun? (not funny? Then cut the project quickly and open a restaurant. Doesn't it mean that the ultimate fate of the failed IT people is to open a restaurant? (is it fun? this question is super serious. This is related to whether your game design can stand the test of users and the market eventually. At this stage, I believe your project management pressure will reach a level, so a reasonable way of development will relieve the pressure when these problems occur.

In the middle of the way, an example of the biggest problem of waterfall development of online games is that some people always say that our game is not complete and does not reflect our real design. Anyway, the core idea is that in the end, he will become a god! It will be a success. I think this is pure bullshit, all kinds of bullshit. When you have deviated from the final direction (high-quality, easy-to-experience game products), the more you do, the more mistakes you make. Of course, the types of games are different, and some do need the integrity of the core gameplay to really reflect the fun of the game. such as casual competitive games.

Well, when your product prototype is in place and the core game has been developed, I suggest turning to agile development. For a specific discussion, please take a look at the decomposition of the problem.

two。 The second problem is to improve, upgrade and repackage the existing product foundation. I recommend that this kind of project starts with an agile development approach.

The advantage of agile development is that it constantly corrects the way forward in order to reach the infinity of the destination (although no one knows where the destination is).

A brief review of what to do, assuming that question 1 continues

First of all, we need to form an organizational structure that can be used for agile development, specifically wikis.

Next, we are going to begin to form our work task option (Backlogs). Since it is an option, there must be a lot of it.

And then, with so many tasks, it's impossible to do it all at once, so let's evaluate the critical path and divide the priority and workload time.

So, a great Backlogs is formed, and we also have to arrange the stage development plan (sprint) according to the schedule of our own project. Pay attention to the decomposition granularity and fullness of man-hours.

OK, everything is ready, start to work, everyone has received their own task, so everyone is crazy to work overtime.

After adding 4 days, on Friday, we will do the integration work, complete the integration, deliver the test, and deliver the acceptance (we can also consider whether to adopt continuous integration according to the progress of the project)

Well, after a month of twists and turns, we made four versions, and finally merged into the X-month beta version, delivering internal and external customer evaluation.

Compared with waterfall development, the delivery cycle is greatly shortened. In waterfall development, it often depends on milestones, but the interval between two milestones is sometimes more than three months or more, which has a huge impact on the rapid response of the project to the market or operation.

Therefore, agile development, in the game project, is more suitable for the later stage, after the completion of the core play function development, the volume and tuning. Or peeling, platform transplantation, expansion, etc.

3. The third question is very interesting. I believe people who use third-party engines will face this problem. You say you have accumulated. Yes, but it's sad that it doesn't have anything to do with what you're going to do. Well, let's also analyze what I think is the right way.

In this case, it must be a mixture of agile and waterfall. How to do it?

First of all, engine shaping, you save a lot of basic work, congratulations.

Then, you need to have an agile phase to determine how to adapt the engine to the new product (you must change it, or determine your new production plan based on your needs).

In addition, you still need to do a waterfall development, do your product prototype, with tools, just simplify the time you make tools, not simplify this piece

Then, when the product prototype is passed, you still need to develop it, and this waterfall is OK.

When your development is over, you have to take the agile route and adjust the product.

Well, after reading this, some people will ask, is there a new development model that unifies them? Of course there is, you can refer to FDD demand-driven development, but I think it is a bit difficult, difficult, he needs to have a very effective development model as a guide (which solves the problem that agile development does not have medium-and long-term accurate goals), but the development model of this thing, that is, large companies have time to sum up, generally small companies and small teams are winning by experience.

The user experience of the game is idealistic, and project management can constrain the process, but can't handle the results, so when you don't have the experience, no matter how good the project management is. On the contrary, when your experience is great, but the balance of time, cost and quality is out of control, it may destroy a great project.

Please remember to read the following sentence carefully

I have described too few conditions for these three types of problems, and I have encountered too many problems in practical work. No matter black cat or white cat, the one who makes money is a good cat, don't worry about the development mode, as long as the project manager is used to it and can manage it well, that kind of cat is fine.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.