目前在一般的電腦遊戲中，尤其是萬人網路連線型態的電腦遊戲，通常將巨大的場景分割成許多不相連通的區段。這樣的做法有其實際上的現實因素。第一、場景的細節太多，無法一次全部載入到記憶體。這個問題在3D的環境裡更為嚴重。第二、這樣的作法可以使得同時間需要同步的線上玩家數量減低，達到自然的區隔，因而減少同步訊息的暴增，減少可能的網路延遲。
針對上述遊戲進行的現況，我們計畫發展一套系統，本系統的主要功能與規格由非同步式檔案輸入/輸出系統、區段預取機制、快取模組所組成。目標在提供玩家一個具有超大型的3D場景、但不延遲且流暢的遊戲經驗。在不減低遊戲流暢性的條件之下，隨著玩家所扮演的角色位置的移動，適時的載入所需的3D場景。
本系統的主要概念是把連續的3D場景（scene）切割成許多區段（sections），將這些區段作為基本的串流單位（streaming units）。每一個區段可能包含了許多靜態幾何資料（static geometry）、動態幾何資料（dynamic geometry）、貼圖（texture）、光線（lights）、物件的資訊、以及音效等等。當玩家所扮演的角色移動於場景時，區段將視需要不斷的被載入，讓玩家感覺整個超大型的場景早已經全部載入記憶體中。

In general, game developers usually divide a huge scene into several disconnected sections, especially in Massively Multi-player Online Role-playing Game (MMORPGs). One reason is that a huge scene may involve massive mesh data and can not be loaded into memory at a time. The other is to reduce the synchronous communications among players, decrease synchronous notifications, and diminish the likelihood of a lag of network.
Due to the predicament mentioned above, we are developing a system, inclusive of an overlapped file I/O subsystem, a pre-fetch scheme, and cache management. Our goal is to provide players with smooth navigation in a massive 3D virtual environment without lags. In our approach, the system will load necessary sections adequately in time without notice of pause on gameplay, which depends on the position of player’s character.
The main concept of our system is that a vast 3D scene has been partitioned into many sections, and these sections are dealt as fundamental streaming units. In each section, there are static geometry, dynamic geometry, textures, lights, and sound effects etc. Sections will be loaded into memory or discard, depending on the character’s position while it walks in the complex scene. Accordingly, players experience that a whole vast 3D scene has been loaded into memory.

1. H. Hoppe, Progressive meshes. In Proceedings of SIGGRAPH’96, pp. 99-108, 1996
2. H. Hoppe, View-dependent refinement of progressive meshes, In Proceedings of SIGGRAPH’97. pp. 189-198, 1997
3. H. Hoppe, Efficient implementation of progressive meshes. In Computers and Graphics 22, pp. 27-36, 1998
4. P. Alliez, M. Desbrun, Progressive Compression for Lossless Transmission of Triangle Meshes. In SIGGRAPH 2001 Conference Proceedings, pp. 195-202, 2001
5. A. Lee, W. Sweldens, P. Schoroder, L. Cowsar, D. Dobkin, MAPS: Multiresolution Adaptive Parameterization of Surfaces. In SIGGRAPH’98 Conference Proceedings, pp. 95-104, 1998
6. G. Turk, Re-tiling Polygonal Surfaces. In SIGGRAPH’92 Conference Proceedings, pp.55-64, 1992
7. W. J. Schroeder, J. A. Zarge, W. E. Lorensen, Decimation of Triangle Meshes. In SIGGRAPH’92 Conference Proceedings, pp. 65-70, 1992
8. P. Lindstrom, V. Pascucci, Visualization of large terrains made easy. In Proceedings IEEE Visualization’01, pp. 363-370, 2001
9. B. Y. Chen, T. Nishita, Multiresolution Streaming Mesh with Shape Preserving and Qos-like Controlling. In Proceedings of ACM 2002 International Conference on 3D Web Technology, pp. 35-42, 2002
10. D. Hearn, M. Pauline Baker, COMPUTER GRAPHICS, C Version. Prentice Hall, section edition, 1997.
11. H. Patterson, G. Gibson, M. Satyanarayanan, A Status Report on Research in Transparent Informed Prefetching. In SIGOPS, Operating Systems Review, pp. 21-34, 1993
12. J. Griffioen, R. Appleton, Reducing File System Latency Using a Predictive Approach. In USENIX’94 Summer Conference Proceedings, pp. 197-207, 1994
13. T. K. Kroeger, D. D. E. Long, Predicting Future File-System Actions from Prior Events. In USENIX’96 Annual Technical Conference Proceedings, pp. 319-328, 1996
14. H. Lei, D. Duchamp, An Analytical Approach to File Prefetching. In USENIX’97 Annual Technical Conference Proceedings, pp. 305-318, 1997
15. Truevision3D, http://www.truevision3d.com