-Google engineers talk about the future of Web

Document directory

• Web applications vs documents, and Web nature
• Program details
• Priorities and permissions of Web Applications
• Extended Chrome OS UI
• Local applications vs. Web Applications
• Built-in Media Player

Http://news.csdn.net/a/20100123/216670.html

 

 

This is an interview between ARS technica reporter Jon stochastand Ryan Paul with Matthew papakipos, a project engineer for Google Chrome, and Eitan bencuya from the Public Relations Department. He talked about the Google Chrome operating system, current Situation and technology: HTML 5 and the Future Web. This is the second part. For the first part, see Google Chrome operating system development.

Name and abbreviation of the interviewee and the interviewee:

• JS-Jon stochs
• RP-Ryan Paul
• MP-Matthew papakipos
• EB-Eitan bencuya

Web applications vs documents, and Web nature

JS: There are two Web modes: Document, Google Maps, and ARS technica, but you put the documents and applications in a browser window at the same time.

I know that you are still improving the user experience. What I want to know is that you will stay in the document mode or switch to the application mode. In this way, your applications will have windows like OS X, rather than the current web page.

MP: Your problem is very good. Today's web people are a little weird. Most of them are static content, which looks more like a document, or even CNN, The New York Times, websites such as Wall Street Journal are still some static Web pages. Although their content is dynamically output, they also contain advertisements. In essence, you read from the pages, unlike Gmail, Picasa, netflix websites are more like applications. You can drag and drop something to change the order.

Currently, in chrome, we have not made much distinction between the two, because Web is not strictly differentiated by the two, but we started to do a lot of things in Chrome OS and added some special features, these functions are real applications. For example, the mailto link in Chrome OS is usually a Windows email program that opens outside the browser, but if you use Gmail, it will be loose. Therefore, we are studying how to directly jump from the current tab to the tab or window running Gmail, or directly pop up a wordpad, you will not interrupt what you are currently browsing.

Therefore, we started to study how to locate different links to different web references. For example, if I have a web JPEG Image Viewer and someone wants to open a JPEG image on a chrome or Chrome OS device, how can I associate it?

There is also an example. DOC file. If I click a DOC file, do I want to open it in office live, use gview to view it, or store it on the U disk? How can a user tell chrome what he wants to do?

JS: This seems to be a problem encountered in early windows, such as Ole and registry. However, on the web, this is equally important, except that the environment is different.

MP: I hope to solve these problems in a simple but flexible way. In Windows, it is very complicated. Different programs want to compete for the processing right of some documents, quickTime and Windows Media Player fight, and the latter fight with chrome, which is difficult for users to handle, and now the Web has begun to have this capability, you can register the file format, we are leading in this regard.

Another question you have is, will we make these applications look like tabs or something else? At present, we just want to make them look like tabs and some windows that don't look like tabs, making them all better.

Most of my time in a day is in a browser. Every time I have to run apps outside of a browser, I feel very troublesome, so we want to keep them all in the tab. However, this all belongs to the UI Layer and will change in the future. Our consumption-oriented release will change many things a year later, but that is our current idea.

JS: Mailplane is very popular in our ARS technicia. Have you used it or have you seen others using it?

MP: No.

JS: At that time, a Gmail packaging application included Mail-related ornaments.

MP: There is something similar in chrome that allows Gmail to run on a special interface. Install a Gmail icon on your desktop, just like a traditional desktop program.

EB: Any website can be opened with a traditional desktop icon.

MP: Google gears allows you to create desktop icons for some applications and make web programs look like desktop programs.

EB: It is built in chrome. Now you can use these shortcuts to open Gmail or any site.

MP:... Do not use those tabs or buttons...

JS: Yes. There is a fluid program that can convert a given URL into a desktop program. Have you ever seen it?

MP: Yes. This is something that he described in chrome already has. I think these are more suitable for traditional operating systems.

JS: I think what I mean by this is that we already have some methods to make Web applications more likeApplicationIn this way, from the perspective of user experience, they are consistent with ordinary applications on the interface, rather than sticking to a web document. This is a good thing for me, better than stuffing everything into the browser window.

But this is a big philosophical topic.

MP: I think we can combine these two concepts in Chrome OS. Now we feel a little weird, because our browsing tabs are managed by browsers, and our operating systems manage different applications. In Chrome OS, everything is tabs. We can have the window manager manage these tabs.

Therefore, in some way, things should be clearer and the user model should be clearer. Because there are no longer two different applications (desktop applications and Web applications-Translator's note), there is only one application. How to demonstrate these will vary from time to time depends on which one is more suitable for users.

Currently, there are two methods to process Chrome OS: Application and page as tab, and preview mode. You can see multiple chrome windows, you can move tabs and Applications in different windows.

However, we are investigating whether these things are called web applications and whether they can have a wider range of capabilities, such as processing file types and mailto links.

Program details

RP: From the programmer's point of view, I am concerned about such a few things. In particular, is there a way to write local code using something like NaCl to implement interfaces between browsers and applications? Does Google consider incorporating things like NaCl into Chrome OS?

MP: Yes. We plan to do this.

RP: Are there any prototypes?

MP: Chrome itself is. You can use some special parameters to run Chrome and call some experimental local customer content.

However, before the release, there will be some changes, not in the stable release of chrome, there are still some security problems, before solving these problems, we do not want to enable these functions by default. But these are definitely in progress and will certainly be included in Chrome OS.

RP: Whether there are plans to expose hardware access to things such as NaCl through JavaScript.

MP: What do you mean by hardware access?

RP: For example, if someone wants to access a camera or GPS, I know that there are APIs like geolocation, but what other peripherals are not supported by web standards?

MP: This is a good question. We are seeking to add those devices to the HTML5 API so that they can be accessed through JavaScript or through C ++ in a local client.

You are right. We still have a lot of work to do. If I want to list those devices, what we need to do now is sound devices, cameras, microphones, line input and so on. In addition, we have no such thing as Dolby's 5.1 channels, P2P networks, or even client-server APIs, we are all doing this.

This is a huge workload, allowing JavaScript and local customers to access these APIs in a few days, but we are doing our best to speed up.

RP: Do you plan to add Dalvik runtime to Chrome OS so that people can run Android programs?

MP: No. We mainly focus on the Web application layer. However, our local customers are a good thing. I am personally excited for it, but our expectations for users should also be pragmatic, and we will be waiting for a while.

The same is true for JavaScript APIs. We hope that JavaScript APIs can access cameras or microphones, but do users have to consider security issues and accept them? How can I prevent some Web tabs from peeking at you through a camera when you are wearing a swimsuit? These problems are tricky.

EB: Of course, we will do it in an open way, but it takes time. HTML5 also takes time.

Priorities and permissions of Web Applications

RP: Can I give some Web applications a high priority?

MP: This is not easy. The current operating system does this. With clear ecls and permission allocation, we learned from traditional operating systems that it is difficult for users to manage them.

Compared with the Web mode, the concept of web is that you should not do anything bad in Web applications. This idea makes the Web run very well. When you browse the Web, in general, there is no need to worry about bad things (the author's note: From the Perspective of applications). Despite browser vulnerabilities and malware, there is no need to worry too much.

We try to avoid ECL and license list problems in traditional operating systems, so we are very cautious, but we are doing our best. Geolocation UI is the most difficult thing we have ever done. This UI is tricky and requires caution because you cannot disclose the geographical location of users, although you can allow users to choose whether to expose their geographic locations to the website, you must make it easier.

If you visit some websites using the HTML5 geolocation API on some mobile phones, a window pops up asking if you are willing to submit your own location. This is annoying. Sometimes, people get used to the First rate. This is very dangerous. We have to figure out how to do it. It's easy to do, but there is no answer yet.

In general, browsers process such things through information bars or dialog boxes, but we tend to use information bars, because the dialog box will lock the browser until you respond, which will reduce the user experience. Because of this, many browsers, including Firefox, tend to use message bars to display prompt messages. You can make a response or continue browsing.

For example, when chrome prompts you to remember the user name and password, the message will be displayed on the message bar, and we tend to be in this mode in terms of license configuration UI issues.

Extended Chrome OS UI

RP: Chrome extension seems to be the main way to extend the UI and operating system, right?

MP: For chrome, this is the case.

RP: Are there plans to add other extension methods for Chrome OS? For example, what features are customized for different platforms?

MP: We haven't thought of it yet. But this is an interesting idea. Do you have a specific idea?

RP: I don't know. I'm just thinking about the menu. I 've added a sidebar to a wide-screen netbook. Although it's useless in a regular browser, however, this may be useful in a browser-themed operating system.

MP: This idea is good. Indeed, when we browse the website, there is a lot of blank space on both sides of the screen. We can place the sidebar in these places, we have already considered the sidebar UI problem in Chrome OS, made some tests, and made constant adjustments to the position of the sidebar. We have over 200 Google testers who use Chrome OS devices every week, we often add new things in it to see if they are often used. Through these experiments, we continue to iterate.

Local applications vs. Web Applications

MP: Ryan, I want to hear your thoughts on Chrome OS. Is there something strange to you?

RP: I have a loose attitude. I think there are many variables. The main problem I think is that compared with chrome browsers based on conventional operating systems, what value does Chrome OS bring to users?

I still have some doubts about incorporating everything into the browser. Let's take a look at some of the current web applications. On the contrary, they are providing some desktop versions, such as Twitter.

I was thinking about how to achieve all these requirements only by using a browser.

MP: Well, if some browsers cannot be implemented, we will expand the browser.

JS: The fluid and mailplane applications we talked about earlier.

MP: I still think that we should look at these applications and think about why they must be local programs. We can add some features for implementation. A good example is that many local programs can run in the background, so people want to implement background processes in web programs, such as uploading files in the background in HTML5, in this way, you do not need to waste a tab and open the browser.

EB: There are also prompt messages. In a Twitter desktop program, whenever you receive a new tweet, you will be given a prompt message. In a browser, there are also notifications API to implement similar functions.

MP: We are still studying the upload problem, because upload is a common use case. When I upload a photo to Picasa, we must open a tab or install some replication programs, we need to work out some ways to build these functions into the web.

Built-in Media Player

MP: Media Player problems. We integrate a complete media player into chrome browsers and Chrome OS. This is often confusing, but subtle, but important, because in a sense, we are integrating something equivalent to a Windows Media Player into chrome.

In normal computers, we need some ways to play JPEG, MP3, and PDF files offline. For example, if you have a USB flash drive with MP3 files on it, you can plug it in, listening to those music, there is no corresponding webpage to control this. We have done a lot of work to process these cases in chrome and Chrome OS.

RP: Your method is standard. Is HTML5-based video and audio?

MP: Yes, absolutely.

RP: Will you enable video tags to support more encoding?

MP: We do not simply add some encoding, but provide a solution for the entire video/audio tag. Because the video/audio tag is designed for webpages, for example, if you have a USB flash drive and want to play an MP3 file on a webpage, what should you do? If you browse those files and play them back.

For example, if you receive an email containing an MP3 file in Gmail, you should not quit Gmail and play the file outside. We hope to play the file directly in chrome, and the whole process should be very clear.

This is our important thing at the UI level. It is not uncommon to integrate the entire media player set framework into a browser, because other browsers have done so, but Chrome has not.

International Source: http://arstechnica.com/business/news/2010/01/chrome-os-interview-1.ars/
Source of Chinese Translation: official site of the comsharp CMS website content management system