【正文】 The buffer solution may also be used with two or more large UI controls sharing the amount of size reduction to be applied. Generalized, this solution ends up as dynamic resizing of the controls in the UI. This may be done using two different approaches. The first is to decide a resizing rule for each window and apply that as tailored code for each window. The second is to have a general layout adjustment algorithm doing it for all windows. Validation The patterns collection was presented at a half day tutorial at the HCI International conference in 2021 [6]. At the tutorial, the structure of the patterns collection was presented, and all patterns were presented at a very brief level. Then, 12 of the 26 patterns were presented in more detail. During the presentation, the participants filled in a questionnaire. They scored the relevance of the main problem areas, the relevance and usefulness of each of the presented patterns, and finally the relevance and usefulness of the patterns collection and such as well as their expectations for future use of the patterns collection. Relevance, usefulness and future use were scored on a scale from 1 (lowest) to 6 (highest). 29 of the participants at the tutorial handed in the questionnaire. There was a small majority of male, age varied from 25 to 50, most being around 30 years old. A majority had their origin in Asia, the rest ing from Europe and America. The majority has an education on master level, the rest split among undergraduates and PhD holders. The educational background was equally split between technical and nontechnical. UI development experience varied from 0 to 12 years, 4 the majority having 5 years or less experience. Experience in developing mobile solutions varied from 0 to 6 years, the majority having 2 years or less experience. Results Scores on the main problem areas show an average score on relevance of for Utilizing screen space, on for Interaction mechanisms, and for Design at large. Scores on the patterns collection as such show an average score on relevance of , and on both usefulness and future use of the patterns collection of . All these scores verify that the patterns collection both addresses relevant problems and gives useful and practical advices on how to solve these problems. Looking at the scores for the individual patterns that were presented in more detail, the scores vary a bit, but are still fairly high. Figure 1 below shows the average scores for relevance and usefulness for the 12 patterns, sorted descending on scores for relevance. Fig. 1. Average scores for relevance and usefulness As for the patterns collection as such, the average scores for relevance are higher than the corresponding scores for usefulness. This is not surprising, as it is usually easier to agree with a problem description than a proposed solution. Although all but one of the average usefulness scores are on the top half of the scale, the patterns where the usefulness scores are lowest, and the patterns where the difference between relevance score and usefulness scores are highest, are candidates for further work. It may also be noted that correlation analyses show that the scores for relevance and usefulness correlates on the level for 7 of the 12 5 problems, and on the level for 2 of the other 5. Also, the scores for relevance and usefulness for the patterns collection as such correlates on the level. Using patterns format to document design knowledge The chosen patterns format is in many ways well suited to document user interface design knowledge, as it captures the essential aspects of a problem. Also, as design patterns may be on different abstraction levels they can be used to describe problems of different “sizes”. Furthermore, dividing a problem field into a limited number of well defined problems makes it possible to handle a set of manageable problems separately. Finally, having a patterns collection makes it possible to bine the just mentioned “divide and rule” principle with having an overall structure. The biggest challenge we had using the patterns format is the connection between problems and solutions. Very often this is a many to many connection. Presenting the same or very similar solutions to a number of problems, either causes a lot of crossreferences or large amounts of repetition. We chose to use crossreferences, as is being done in other patterns collections [2]. Currently we are considering restructuring the patterns collection so that each pattern represents either one solution or a unique bination of one problem and one solution. Both these approaches will reduce the need for crossreferencing, but both will increase the number of patterns, thus making it more difficult to get an overview of the patterns collection. Related work There are a number of patterns collections and even collections of patterns collections on the web, see also [1] for an assessment of such collections. There are also a few collections of patterns for mobile user interfaces, like The Design Pattern Wiki and Little Springs mobile UI design patterns. The latter overlaps with two of our main problem areas, but while our collection is anized by problems, this collection is anized by solutions. The patterns presented in [8], although focusing on mobile interaction, are much wider in its scope than our collection, with only two user interface patterns. [2 amp。 Landay J A (2021) Design Patterns, Course documentation [3] Landay J A amp。 參考文獻 [1] Deng J et al (2021) Managing UI pattern collections. In Proceedings of the 6th ACM SIGCHI New Zealand chapter39。 致謝： 本文是基于工作是由挪威研究理事 會，并在這些項目的行業(yè)合作伙伴資助的本影和 FLAMINCO項目的支持。主要職業(yè)能力分為結構設置一個管理的問題，主要是相同的解決方案可能適用于一些問題，引起了很多模式之間的交叉引用。它還確定了需要更多的工作模式。 已通過驗證的模式集合使用教程問卷。 結論和未來計劃的工作 在本文中，我們已經介紹了移動應用的用戶界面設計模式的