Performance/Meetings/2009 01
Performance/Meetings/2009/01/09
Meeting Minutes
IRC Meeting of Sun Microsystems (StarOffice) with RedFlag2000
Performance Project
Date: 2009/01/09
Time: 16:29– 15:29
Meeting No.:
Agenda:
(4:29:59 PM) Matthias: Hi, all; good morning / afternoon; happy new year
(4:30:01 PM) yugq: Happy new year to all!
(4:30:01 PM) xiuzhi: hi all
(4:30:16 PM) xiuzhi: JackieSun: Are you Sui yinan?
(4:30:25 PM) JackieSun: xiuzhi ,yes
(4:30:25 PM) yugq: JackieSun, hi!
(4:30:31 PM) liheng: Hi,all!
(4:30:35 PM) odf-mib: Hi all!
(4:30:36 PM) JackieSun: I am Sun Yinan
(4:30:39 PM) liheng: Happy new year!
(4:31:16 PM) Dieter_: Hi all :-)
(4:31:18 PM) xiuzhi: Hi all
(4:31:50 PM) Matthias: Hi liheng, xiuzhi, all
(4:31:54 PM) FrankL: Hi, all
(4:31:56 PM) JackieSun: Let's begin?
(4:32:14 PM) Dieter_: Yes, let's start. Malte is on vacation.
(4:32:15 PM) xiuzhi: JackieSun: Please follow the agenda
(4:32:24 PM) JackieSun: ok
(4:33:27 PM) liheng: 1. Plan to work out first User Experience Index for OOo
(4:33:27 PM) liheng:2. When we save a file in ODF, the pictures can not be compressed to save more space as we
(4:33:27 PM) liheng: usually insert image that it has been in compressing-format (like .png & .jpeg) . So can we
(4:33:27 PM) liheng: make a new method for save .od* files, that it do only pack the media files in archive but not
(4:33:27 PM) liheng: compress them. So that we can save more time from compressing pictures,video,audio in saving
(4:33:27 PM) liheng: and uncompressing them in loading.
(4:34:36 PM) liheng: Last IRC, We talk about the Performance server, but Malte and odf-mib think we can do a report in CH2000 first
(4:35:32 PM) liheng: I am very greet. So we can plan to work out first report show User Experience Index for OOo
(4:36:04 PM) xiuzhi: And AnsyLoading reviewer candidate.
(4:36:30 PM) odf-mib: liheng: I have to apologize, but it is not clear to me to which report you are referring.
(4:36:45 PM) liheng: We must decide which version we evaluate,and which hotspots we focus
(4:37:18 PM) Matthias: xiuzhi: can you explain what "AnsyLoading" is / means, please ?
(4:38:12 PM) odf-mib: MHU: It's related to the 2nd agenda item. I suggest we clarify this when we reach it.
(4:38:25 PM) xiuzhi: mhu: for big document. to speed up the display
(4:38:31 PM) liheng: odf-mib:In last mail we talk about do benchmark testing automically, this report is the benchmark report and counting the User
(4:38:54 PM) liheng: ... User Experience Index for OOo
(4:39:44 PM) Dieter_: I think it would be good to have a short introduction of our new team members Frank (FrankL) and Stephan (sb):
(4:39:46 PM) Dieter_: Frank is from User Experience and will work on all the UX stuff and Stephan is from development and will focus on startup performance.
(4:39:55 PM) xiuzhi: mhu:IBM are going to open, but not be available until now. we are developing it by ourselves. because IBM would not to show us their source
(4:40:40 PM) Matthias: xiuzhi: thanks for your explanations [I know "Loading" but have never heard the term "Ansy"]
(4:40:46 PM) xiuzhi: And Dr. JackieSun ,His major is image processing
(4:40:54 PM) liheng: FrankL:Welcome!
(4:41:07 PM) FrankL: Hi
(4:41:07 PM) liheng: sb_:Thanks :)
(4:41:16 PM) sb_: hi folks
(4:41:30 PM) xiuzhi: mhu:using two threads to load.
(4:41:53 PM) Matthias: yes, welcome Frank, Stephan, and Sui Yinan
(4:42:15 PM) JackieSun: I find in MS Office, the size of saved image is depent on the size of rect which the image is inserted, the more big of the rect , the more size of saved image
(4:42:17 PM) liheng: odf-mib: And the User Experience Index is only one number to show whole performance of OOo for all developers and users
(4:42:52 PM) xiuzhi: JackieSun: a minute. the second agenda item, please waiting
(4:42:59 PM) liangjun: welcome Frank, Stephan, and Sui Yinan:)
(4:43:03 PM) JackieSun: ok
(4:43:12 PM) JackieSun: 3ks everyone
(4:43:49 PM) liheng: Which version we do first report with?
(4:44:42 PM) liheng: Xiuzhi and me think 3.0 is good for it
(4:45:02 PM) Matthias: yes, I think 3.0 is the most interesting nowadays
(4:45:16 PM) Dieter_: +1
(4:46:23 PM) liheng: We will start to choose some case and make a formula for counting the Index, and publish them next week
(4:46:46 PM) Matthias: sounds good
(4:47:42 PM) liheng: then we will do whole report in 2 weeks
(4:49:21 PM) liheng: Startup and Load/Save is mainly hotspots for first benchmark testing report
(4:49:48 PM) Dieter_: would be good to come up with a draft report first so that Frank could jump in an work with you
(4:50:28 PM) liheng: Dieter_: OK
(4:50:48 PM) Dieter_: liheng: thanks :-)
(4:51:28 PM) liheng: We will get the report at Jan,31th
(4:52:54 PM) liheng: All:Can we turn to the next?
(4:53:50 PM) xiuzhi: odf-mib: What is your idea?
(4:54:11 PM) odf-mib: About what?
(4:54:20 PM) liheng: When we save a file in ODF, the pictures can not be compressed to save more space as we
(4:54:20 PM) liheng: usually insert image that it has been in compressing-format (like .png & .jpeg) . So can we
(4:54:20 PM) liheng: make a new method for save .od* files, that it do only pack the media files in archive but not
(4:54:20 PM) liheng: compress them. So that we can save more time from compressing pictures,video,audio in saving
(4:54:20 PM) liheng: and uncompressing them in loading.
(4:54:51 PM) xiuzhi: odf-mib:the second item.
(4:55:11 PM) odf-mib: Okay. Let me check if I do understand it:
(4:55:35 PM) JackieSun: I find in MS Office, the size of saved image is depent on the size of rect which the image is inserted, the more big of the rect , the more size of saved image
(4:56:39 PM) odf-mib: Or, let's start with a question: Is the idea to
(4:56:55 PM) xiuzhi: JackieSun: two aspects. one is Liheng mentioned. the other is yours
(4:56:58 PM) odf-mib: a) save the files uncompressed,
(4:56:59 PM) Matthias: JackieSun: is there in (MSO) an option to save image in original size (instead of default actual used size) ?
(4:57:12 PM) odf-mib: b) to save them with a reduced resolution?
(4:57:13 PM) JackieSun: mhu, i think no
(4:57:28 PM) Matthias: JackieSun: okay, thanks
(4:57:44 PM) yugq: The .png or .jpg files can't be compressed in size anymore by zip format. But if we try to compress them, it just waste time to do nothing.
(4:58:09 PM) liheng: odf-mib:we compress all XML file and bind all media files behind the ZIP
(4:59:25 PM) liheng: So we can save time of compressing and uncompressing media files when load and save a ODF file
(4:59:26 PM) odf-mib: linheng: I'm not sure. I didn't check this but my understanding is that we store files that are already compressed (like JPEG or PNG) uncompressed in the zip file.
(5:00:03 PM) JackieSun: I agree with odf-mib
(5:00:35 PM) Matthias: odf-mib: I think, we have this "presentation optimizer" ? can we combine this somehow (to be defined) so that we can have this feature, and don't loose image resolution if the user wants so?
(5:00:55 PM) liheng: I mean the media files are not in ZIP:)
(5:01:18 PM) odf-mib: Sorry, I'm still not sure we all talk about the same.
(5:01:58 PM) odf-mib: We are talking about compression, and we are talking about the resolution of the image size.
(5:02:07 PM) Matthias: yes, both
(5:02:16 PM) liheng: * ZIPPED *
(5:02:16 PM) liheng: * XML files *
(5:02:16 PM) liheng: *Media files*
(5:02:29 PM) JackieSun: the media files is not in ZIP and the jpeg file is in, is it ?
(5:03:33 PM) xiuzhi: all of the image and media files are out of the Zip
(5:03:44 PM) xiuzhi: That is LiHeng's mean
(5:03:48 PM) yugq: When save a doc, before zip all files into a odf, the media files like pictures is compressed format already. So compressed pictures don't need to compress a second time.
(5:04:53 PM) odf-mib: yugc: I strongly believe they aren't compressed a 2nd time.
(5:05:30 PM) liheng: odf-mib:but they are in the Zip file
(5:05:30 PM) odf-mib: yugc: ZIP allows you to store files uncompressed. We use that feature for meta.xml, and I strongly assume also for images.
(5:05:42 PM) odf-mib: yugc: If not, I would consider this to be a bug.
(5:06:51 PM) odf-mib: liheng: Yes, they are in the zip file. So it may require some more time to save them on disk when a presentation is stored. That's is saving time. But not time required for compression.
(5:07:55 PM) odf-mib: liheng: One can also just reference a image in Impress rather than embedding it. In this case, one the link to the image is stored, but not the image itself.
(5:08:17 PM) odf-mib: liheng: Is something like this what you have in mind?
(5:09:12 PM) yugq: odf-mib: But in fact, the .jpg & .png like files is a compressed format. When we compress a jpg file to a zip, the size would be close. But zip a file comsume time. So we could not zip a jpg file.
(5:09:43 PM) FrankL: Most times graphic keep untouched in a document. But when saving a document they will be stored again and again. Could this be avoided to save time?
(5:10:12 PM) xiuzhi: We want to save the saving time when the image is embeded , how to do that?
(5:10:35 PM) odf-mib: yugc: Zip allows you to store files without compression. In this case, no time is required for compression. Only the data needs to be copied into the file.
(5:10:45 PM) liheng: yugq:images in zip is same file as it original size
(5:11:50 PM) yugq: odf-mib, yes. I got you now. Thank you!:)
(5:12:12 PM) liheng: odf-mib:I think image can save out of ZIP and in a mapping format.
(5:13:26 PM) JackieSun: can the performance be increase by not saving the jpg files in a ZIP?
(5:14:05 PM) odf-mib: liheng, JackieSun: This option does exist already. Image can be inserted by a reference, rather than being embedded.
(5:14:31 PM) odf-mib: The issue with this is the user experience than copying file from one place to another.
(5:15:20 PM) JackieSun: OK
(5:15:39 PM) FrankL: Could we handle graphics which are already inside the doc zip as a reference? Then we do not have to store them again and again.
(5:15:39 PM) odf-mib: As a side note: I've just dumped a file containing an image. Here is an extract:
(5:15:54 PM) odf-mib: 11123 Defl:N 1708 85% 01-08-09 12:15 842340f8 Pictures/200000B100000FE500000EF9FF04DDF4.svm
(5:15:57 PM) odf-mib: 103 Stored 103 0% 01-08-09 12:15 d0f5f678 Pictures/100002000000000A0000000ADDA84F49.gif
(5:16:33 PM) odf-mib: So, jpg's are just stored, while SVM files (which are not compressed by itself) get compressed by ZIP.
(5:16:37 PM) liheng: odf-mib:Yes, so we can package image behind ZIP part
(5:17:39 PM) liheng: I think we need not compress and media file to reduce size of file.
(5:18:31 PM) xiuzhi: now odf = zip(xmles+ images+medias). Can it be :odf=one zip file(zipped xmles) + images+medias. ? Is this your mean LiHeng?
(5:18:55 PM) sb_: liheng: add data to the end of a zip file, after the data that follows the zip protocol? would that not break standard zip tools reading/modifying odf files? is it worth that?
(5:18:57 PM) liheng: xiuzhi:Yes!
(5:20:21 PM) odf-mib: liheng: Would the images be separate files, or appened to the zip as _sb assumes?
(5:20:24 PM) liheng: sb_:Yes it is a problem i worried,but it can save time when we loading and saving a document.
(5:20:59 PM) sb_: liheng: can it save so much time that we should even consider it?
(5:21:32 PM) JackieSun: Liheng: I think it is not a good idea. When a user want copy them to another PC, he should select some files but not one, isn't it?
(5:21:32 PM) liheng: I think so.
(5:21:50 PM) Matthias: liheng: how can it save time? you need to load the image anyway, either from zip-container or from disk ?
(5:23:07 PM) Matthias: ...and yes, with an external reference, a user would not easily know what to copy (it is not one file)
(5:23:20 PM) odf-mib: mhu: For loading, you are right. And I think we can even not save time while saving. It should not make a difference whether we store uncompressed data in the zip file, or append it to it. The amount of data that needs to be stored is the same in both cases.
(5:23:25 PM) liheng: Matthias:we can gather all media file in a section.
(5:24:06 PM) Matthias: liheng: I dont understand what you mean, can you explain ?
(5:24:19 PM) odf-mib: liheng, all: Before we discuss the options, I would like to understand if we want to save time while loading, or while saving a document.
(5:25:07 PM) liheng: Matthias:I send some details to you all after this meeting,:)
(5:25:35 PM) liheng: To explain how to save some time
(5:25:43 PM) Matthias: liheng: okay, yes we can discuss this offline
(5:25:51 PM) xiuzhi: +1
(5:26:00 PM) liangjun: :)
(5:26:06 PM) liheng: Yes, I will send it to performance mailing list
(5:26:21 PM) liheng: odf-mib:Thank you!
(5:26:29 PM) Matthias: good, so everyone can participate
(5:26:45 PM) sb_: liheng: performance mailing list = performance@tools.openoffice.org?
(5:26:54 PM) Matthias: sb_: yes
(5:26:56 PM) liheng: sb_: Yes
(5:27:14 PM) liheng: sb_:Thanks for your attending again
(5:27:17 PM) liheng: :)
(5:27:23 PM) JackieSun: ok, let's discuss how to deal with the high-res images next time
(5:27:42 PM) Matthias: okay
(5:27:54 PM) liheng: OK, that all today for me
(5:27:58 PM) xiuzhi: JackieSun: You can email your idea to the performance mailinglist
(5:28:08 PM) JackieSun: xiuzhi: ok
(5:28:21 PM) Matthias: yes, that would probably be good
(5:28:45 PM) Matthias: okay, then see you next week; bye all
(5:28:56 PM) yugq: bye all!
(5:28:58 PM) JackieSun: bye all
(5:29:01 PM) odf-mib: by all!
(5:29:03 PM) liheng: All:see you next time.
(5:29:03 PM) liangjun: bye all!
(5:29:12 PM) FrankL: bye!
(5:29:16 PM) sb_: bye
(5:29:51 PM) xiuzhi: bye
