|
|
http://pvanhoof.be/blog/index.php/2005/06/19/38-desktop-integration-tomorrow
With this blog-entry I'm probably about to make one of my major mistakes in my career. I'm probably going to make a lot people pissed about this. And that's okay. A blog is about sharing your opinion no? This is my opinion. Enjoy it.
Last night, I dreamt about Utopia. I dreamt that some day I would copy-paste a part of a video to my E-mail. And while I was typing it, my E-mail client was playing what I've just copy-pasted. I dreamt that the video wasn't transferred from the clipboard-owner to my xserver and back to the clipboard-requester. You see, my xserver runs remote! Whereas both the clipboard-owner and requester run on the same machine. In my dream, the copy-paste was a local transfer of clipboard-data between two processes sharing the same memory, kernel and CPU. In fact I even dreamt that the process originally owning the video-data was the one playing it on the canvas of my E-mail client during the time I was typing the E-mail. And that only after I decided to actually send the E-mail, the data got transferred.
I dreamt that after that copy-paste operation, I copied the video-fragment from my E-mail editor to my instant messaging application. I wanted to show my friend what I was about to E-mail to the recipients of the E-mail. I copy-pasted it with some marked up text and it just got pasted perfectly into the messenger window. And my friend really received it just like how I wanted him to receive it. And his messenger client started playing the video while it was getting transferred. And our chat wasn't blocked.
I dreamt that during the video-transfer I was chatting in combination with instant video conferencing. I dreamt about integration of GStreamer with X11's clipboard. I dreamt about applications really working together to achieve very typical and common desktop usage. I dreamt that there wasn't any difference between an application that has it's name starting with a K and an application that has it's name starting with a G. For the end-users it just didn't matter at all! My instant messenger was a K-one. And my E-mail client was .. whatever .. Evolution.
I dreamt that all configuration information was shared using a common system. And as the end-user of my system, I never ever cared about this configuration. At my company our administrators took care of all the configuration stuff from their seat. And both applications that have their names starting with a K and the ones with a G just worked. You know like .. they just worked! And it just didn't matter, for the administrator, how the configuration management of those applications is done.
Using some tool they decided about company-wide configuration options. When they decided to change one, not a single user had to restart any of the applications. The configuration-change was fully event-based for the applications.
And everything just worked together with my personal data assistant, with my cellphone, with my Nokia 770 or whatever version of the device I was dreaming about.
But then I woke up. I realized there isn't any organisation trying to persuade free software developers to use some standardisation nor to decide on important shared desktop components. Oh you have freedesktop.org. But when asked they just say: "Hey, how!! We don't do standardisation! We just give free software developers a forum to start flaming^H^H^H^H^H^H discussions about standardizations!".
And I realized there's a massive amount of political bullshit about technologies in the free software world. I realized that stuff like, integrating the X11 clipboard with a technology-library like GStreamer is probably a typical "over my dead body" thing for a huge amount of free software developers in our community. I realized that the current components for storing configuration data just aren't shared between the KDE and GNOME platform. And that getting there is politically-wise so damned hard. You just couldn't believe it. I'm serious! Can you imagine that this little component causes "over my dead body"-type reactions on mailinglists? It does. Again, I'm serious.
Yet during GUADEC some Novell people tried to tell us: "Jow dudes! Stop doing infrastructure-work. Start finishing real applications, we have good infrastructure already!".
I'm just ... not sure about that. Our infrastructure lacks a lot integration possibilities. Like infrastructure for integration between KDE and GNOME applications. Our basic infrastructure often isn't consistent between the KDE and GNOME developer platforms. There's progress! I agree with that statement. Like the fact that GStreamer might become a shared multimedia library for both. Or like the fact that D-BUS appears to be liked by both worlds. That's good. I'm thrilled about it (I'm being honest)! But I'd share your opinion if you say that we need more like this. I don't think we already have a kick-ass developer platform. We don't.
For a person who has a huge amount of free-time to learn about all the quirks of our platforms, we have more or less two platforms that are almost good enough. They can't yet work together. But that's okay -- no it's not!! --. I don't think a lot software houses will really already like our developer platform. Or will like the fact that they'll have to choose between letting the name of the application start with a K or a G. You see, that letter isn't the real problem...
The real problem is the fact that it will have a great impact on their development model. On the type of persons to employ. On the libraries being used. On the many incompatibilities with whatever distribution ships with the other major desktop platform as default. Are we really so naive to believe that such software-building companies will let their customers change the major default setting of a distribution that has expensive support? -- Just to give one example why not to do this --
How is a salesman going to sell (custom) Linux desktop software? Does he actually have to ask the customer: "Do you run KDE or GNOME". And depending on that answer tell the customer: "Oh I'm sorry, we don't have developers for that platform, we can't help you". Has this free software community ever professionally thought about the impact of that stupid little question? Have we ever professionally thought about that this might be the thing that is holding back wide adoption of the free desktop platform? Have we ever considered really talking with each-other about this? Not just giving one K-developer some beer at GUADEC and persuading him to adapt his kick-ass K-infrastructure. I mean, really talking. Really getting somewhere. Really making those important decisions. Nevertheless I agree that inviting them is a very, very very very good idea and nobel thing to do. I'm not saying it wont change stuff. I'm saying that I fear that it's not enough.
For those people who are planning to respond: "Hey moron, in stead of whining you should start developing and fixing what you hate so much": I am one of the persons who is at this moment spending a huge amount of his free time on the architecture and implementation of DConf. Thank you, come again.
感觉(大部分桌面功能) 2000年的Beos 好过如今的 linuxdesktop |
|
IP卡
狗仔卡
发表于 2005-6-20 13:46:40
提升卡
置顶卡
沉默卡
喧嚣卡
变色卡
显身卡
发表于 2005-6-20 15:47:36
