Ok, I've had some time to digest the two packed days of wineconf. It was great being around so many smart and passionate people. I've long since given up coding, so much of it went over my head but the various biz presentations hit home and I had a chance to talk to many of you to crystallize some thoughts. We need 20 million avid Linux desktoppers. Look at what Apple's vocal minority gets them. With 4% marketshare they command sections in stores. They get hardware with Mac/Win drivers. They get documentation in the box. Even stores dedicated just to them. That's what we need to get Linux really going. If we can get the user inertia and the business inertia around Linux, then that will take it to new heights. I believe Wine will play a role in this. But I think a slight re-focusing may be necessary. Have you ever tried to learn how to play the guitar? I think most people have at one time. First you start learning chords and you say to yourself "I'm going to learn all the chords so I can play all the songs". After about struggling through 4 of the 700 chords, you realize that it will take until 2037 to learn all the chords. To make guitar rewarding, you have to start with a list of popular songs and learn the chords just for those. This way you get that feeling of success and it motivates you to stick with it, even if it's just learning Happy Birthday. I believe a similar strategy could really energize Wine. We need a "top 10 tree". What WINE needs is a specific value promise to the consumer. In many ways, the same challenge confronts WINE that confronts Linux. Great things will happen *if* we get a bunch more users. But to get more users, there needs to be some specific value that people get from Wine. And it can't be fleeting value. It needs to allow them to do something specific and when they get the new version, it needs to no take away that value, but add more. The challenge we have now is that the goal for WINE is an architectual one (learn every chord). While nobody can deny the value of a solid underpinning, users don't care about the architecture. They want to hear music! And with Wine this means "what programs can I run?" I would also suggest that focusing around specific applications focuses energies of developers on solving specific user problems. The more problems you solve, the more users will get excited about Wine. One question arose at wineconf which never got addressed. "How do we get more people excited about Wine?" I've been thinking about that. I believe a specific part of the answer is making Wine actually work for a subset of programs. Take it from a theoretical white paper stage to a stage where people actually use it. I'm not suggesting its at a white paper stage, but rather if the world can't use it with any regularity for even a narrow set of applications, the perception is it's nothing more than theory. What I believe needs to happen is that we have a 'top 10' tree. A version of Wine for which the primary goal is to do a good job of running a set version of win32 programs. This serves both parties. It gets developers all laser-focused on the same goal. Because we've narrowed the universe we can do specific testing. And perhaps MOST important, consumers know what to expect. If you tell them "we have a Linux OS which will run Win32 programs, they'll bring out a geneology program from 1992 which won't run and then they'll think it sucks. If however the public commitment is "this software is designed to run the following: MS Word Lotus Notes Quicken" and it actually works, you have a believer. You have someone who will get excited and will offer their help. You give momentum to Wine. You get program sponsors (those willing to oversee ongoing supervision/testing of programs. You get more people offering to be on documentation teams. And you get more developers. One final thought. I think the only tenable way to implement a Top 10 type approach is to get widespread support. Lindows.com alone is too small to do this in a timely fashion. So is Codeweavers, transgaming, etc. The only approach with any chance of success is one which leverages the entire community. I can assure you that Lindows.com would put our full weight behind such an endeavor. While we only have a couple of coders, we've got a super QA department. We've got solid organizational skills. We have an enthusiastic user base we can tap for testing and support. I'll shutup now before I get banned from wine-devel. -- MR Ok, those are my thoughts.
"this software is designed to run the following: MS Word Lotus Notes Quicken"
You really have two tasks at hand. 1. Running some windows apps 2. Moving People off of Windows Apps To reach these goals you will only need to support a limited number of software configs. I will propose my list with some suggestions. Your top 10 list should be something like this and during the install or post install you should offer the user the option to "Upgrade" these apps. 1. Office (Word, Excel, PowerPoint, Outlook.) 2. Internet Explorer. 3. AOL 4. TuboTax 5. Quicken/QuickBooks 6. Lotus Notes 7. All of the P2P software 8. Juno 9. Act 10. The Palm Desktop 1. StarOffice/KOffice (Bitch because of no Outlook clone) 2. Mozilla/Konq (Auto Import of Favorites?) 3. Wine+AOL (Intergrate with Mozilla) 4. Wine+TuboTax/Moz? 5. Wine+Quicken/QuickBooks? 6. Wine+Lotus Notes 7. Wine+All of the P2P software/Native P2P Software 8. Wine+Juno/Moz? 9. Wine+Act 10. Wine+The Palm Desktop IMHO if this is the type of plan you have in mind. You should set your development team to do regression testing on every api these programs use...I dont know how much more the wine developers can do for you but I am sure that if they break something that should already work, they will try to fix it. On a Side Note: One thing that Wine/Lindows and Reactos will need is to be able to run alot of the newer programs with tighter Mozzila/NS intigration. This was discussed a little during James's presentation on the Jscript.dll interface to Seamonkey. Some of the newer software loads Internet Explorers MSHTML.DLL for rendering and I belive that this software with a little help from wine could run better with tigher Mozilla Intergration. Thanks Steven Edwards Think Outside the Box. __________________________________________________ Do You Yahoo!? Yahoo! Sports - live college hoops coverage http://sports.yahoo.com/
1. StarOffice/KOffice (Bitch because of no Outlook clone)
I think that Ximian Evolution gets pretty close to being a Outlook clone. They even have a product (not free :( ) that allows it to connect to an Exchange 2000 server. http://www.ximian.com/products/ximian_evolution/
On a Side Note: One thing that Wine/Lindows and Reactos will need is to be able to run alot of the newer programs with tighter Mozzila/NS intigration. This was discussed a little during James's presentation on the Jscript.dll interface to Seamonkey. Some of the newer software loads Internet Explorers MSHTML.DLL for rendering and I belive that this software with a little help from wine could run better with tigher Mozilla Intergration.
As a starting place for this, developers will want to visit: http://www.iol.ie/~locka/mozilla/mozilla.htm http://www.iol.ie/~locka/mozilla/control.htm This contains the work on making Mozilla a plug-in replacement for IE as a control on windows (implementing IBrowser, IBrowser2, with a simple DOM implementing IHtmlDocument2, IHtmlElement, IHtmlElementCollection, etc.) The possible work involved would be: (A) making sure this control is as feature complete as possible and (B) getting WINE to run Mozilla properly. Both of these tasks are non-trivial. If instead someone was planning to try to interface the native Linux Mozilla binary with WINE, you should realize that would be even more difficult, one immediate showstopper is that Mozilla uses pthreads and WINE doesn't. It would be a nightmare. -James -- James Hatheway james(a)macadamian.com ~ http://www.macadamian.com Software Designer - Macadamian Technologies, Inc. -- Software experts for the world's leading technology companies. Obligatory Geek Expression: "Save the Whales, Free the mallocs!"
I've been working on trying to get the embedded browser in AOL to work, even WITH the native mshtml.dll, but it won't. My concern is that if the infrastructure isn't even there to create the com object, it might be a lot of work to get mozilla to launch. I plan to submit patches that allow aol7 to run today, along with some detailed info on the browser problem, but if you have any suggestions as to how to make the embedded browser work, that would be great. James Hatheway wrote:
1. StarOffice/KOffice (Bitch because of no Outlook clone)
I think that Ximian Evolution gets pretty close to being a Outlook clone. They even have a product (not free :( ) that allows it to connect to an Exchange 2000 server. http://www.ximian.com/products/ximian_evolution/
On a Side Note: One thing that Wine/Lindows and Reactos will need is to be able to run alot of the newer programs with tighter Mozzila/NS intigration. This was discussed a little during James's presentation on the Jscript.dll interface to Seamonkey. Some of the newer software loads Internet Explorers MSHTML.DLL for rendering and I belive that this software with a little help from wine could run better with tigher Mozilla Intergration.
As a starting place for this, developers will want to visit: http://www.iol.ie/~locka/mozilla/mozilla.htm http://www.iol.ie/~locka/mozilla/control.htm This contains the work on making Mozilla a plug-in replacement for IE as a control on windows (implementing IBrowser, IBrowser2, with a simple DOM implementing IHtmlDocument2, IHtmlElement, IHtmlElementCollection, etc.) The possible work involved would be: (A) making sure this control is as feature complete as possible and (B) getting WINE to run Mozilla properly. Both of these tasks are non-trivial. If instead someone was planning to try to interface the native Linux Mozilla binary with WINE, you should realize that would be even more difficult, one immediate showstopper is that Mozilla uses pthreads and WINE doesn't. It would be a nightmare.
-James
I've been working on trying to get the embedded browser in AOL to work, even WITH the native mshtml.dll, but it won't. My concern is that if the infrastructure isn't even there to create the com object, it might be a lot of work to get mozilla to launch.
One thing that could be snagging you, is that make sure you are also using native oleaut32.dll. The implementation of type library loading, and IDispatch interface implementation (especially IDispatch::Invoke) inside WINE is incomplete, and definitely not stable enough to run a non-trivial app like AOL. We (Macadamian, mainly Francois) did a lot of work in this regards, but some of it didn't get into the tree because it was too hackish. The patch should be in the archives somewhere. -James -- James Hatheway james(a)macadamian.com ~ http://www.macadamian.com Software Designer - Macadamian Technologies, Inc. -- Software experts for the world's leading technology companies. Obligatory Geek Expression: "Save the Whales, Free the mallocs!"
If instead someone was planning to try to interface the native Linux Mozilla binary with WINE, you should realize that would be even more difficult, one immediate showstopper is that Mozilla uses pthreads and WINE doesn't. It would be a nightmare.
.....I didn't really think about that. What about applications ported with winelib, can you use pthreads with them? If such a wrapper was developed to run Mozilla_W32/Wine could you then use winelib to port the application and use linux_mozzila/gecko? Steven _________________________________________________________ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com
"Steven Edwards" <Steven_Ed4153(a)yahoo.com> wrote:
If instead someone was planning to try to interface the native Linux Mozilla binary with WINE, you should realize that would be even more difficult, one immediate showstopper is that Mozilla uses pthreads and WINE doesn't. It would be a nightmare.
.....I didn't really think about that. What about applications ported with winelib, can you use pthreads with them? If such a wrapper was developed to run Mozilla_W32/Wine could you then use winelib to port the application and use linux_mozzila/gecko?
I don't know the status of the patch, but Transgaming did some work to use pthreads natively in WineX. -- Dmitry.
.....I didn't really think about that. What about applications ported with winelib, can you use pthreads with them? If such a wrapper was developed to run Mozilla_W32/Wine could you then use winelib to port the application and use linux_mozzila/gecko?
I don't know what Winelib supports, have never played with them. In any case, no matter which approach you take it will be tricky and time consuming. I have a feeling going linux mozilla native would be more painful, but I can't back that up with actual investigation. Just a hunch. -James -- James Hatheway james(a)macadamian.com ~ http://www.macadamian.com Software Designer - Macadamian Technologies, Inc. -- Software experts for the world's leading technology companies. Obligatory Geek Expression: "Save the Whales, Free the mallocs!"
On Wed, Mar 20, 2002 at 10:35:49AM -0500, James Hatheway wrote:
are non-trivial. If instead someone was planning to try to interface the native Linux Mozilla binary with WINE, you should realize that would be even more difficult, one immediate showstopper is that Mozilla uses pthreads and WINE doesn't. It would be a nightmare.
In a monolith world, yes. The "unix way(tm)" to do it is to separate process spaces. Still not a small task. regards, Göran
In a monolith world, yes. The "unix way(tm)" to do it is to separate process spaces.
Uhm, but in windows this is all (or at least can be) done in-proc. So at least the glue code must be inproc. I suppose your glue could handle marshalling all the variants to your out-of-proc linux process.
Still not a small task. True.
-James -- James Hatheway james(a)macadamian.com ~ http://www.macadamian.com Software Designer - Macadamian Technologies, Inc. -- Software experts for the world's leading technology companies. Obligatory Geek Expression: "Save the Whales, Free the mallocs!"
On Thu, Mar 21, 2002 at 04:14:20PM -0500, James Hatheway wrote:
In a monolith world, yes. The "unix way(tm)" to do it is to separate process spaces.
Uhm, but in windows this is all (or at least can be) done in-proc. So at least the glue code must be inproc. I suppose your glue could handle marshalling all the variants to your out-of-proc linux process.
Correct. Since wine now is LGPL, if I should take on this task I would look at orbit/bonobo from the Gnome project and how the mozilla component in nautilus works. If there is some usable in there a generic COM object with glue to bridge Wine/COM to bonobo would be a nice addition to a wine based desktop.
Still not a small task. True. Very true
regards, -- Göran Thyni
At 01:14 AM 3/20/02 -0800, Steven Edwards wrote:
"this software is designed to run the following: MS Word Lotus Notes Quicken"
You really have two tasks at hand. 1. Running some windows apps 2. Moving People off of Windows Apps
I don't see number 2 as a task. Our goal is to just implement task 1. Task 2 has nothing to do with WINE primarily. There are some Windows Apps that have no Linux equivalent, think of all the games...
1. Office (Word, Excel, PowerPoint, Outlook.) 2. Internet Explorer. 3. AOL 4. TuboTax 5. Quicken/QuickBooks 6. Lotus Notes 7. All of the P2P software 8. Juno 9. Act 10. The Palm Desktop
Plus tons of games Roland
1. Office (Word, Excel, PowerPoint, Outlook.) 2. Internet Explorer. 3. AOL 4. TuboTax 5. Quicken/QuickBooks 6. Lotus Notes 7. All of the P2P software 8. Juno 9. Act 10. The Palm Desktop
Plus tons of games
I forgot to add chat clients. Move everyone from AIM/MSN/ICQ to Trillian under wine or everybuddy/GAIM __________________________________________________ Do You Yahoo!? Yahoo! Sports - live college hoops coverage http://sports.yahoo.com/
On Wed, 20 Mar 2002, Steven Edwards wrote:
I forgot to add chat clients. Move everyone from AIM/MSN/ICQ to Trillian under wine or everybuddy/GAIM
I think they're not that important because good native Linux applications (licq, gaim, everybuddy, kicq, ...) exist already. LLaP bero -- This message is provided to you under the terms outlined at http://www.bero.org/terms.html
On March 20, 2002 01:47 am, Michael Robertson wrote: [...]
they'll think it sucks. If however the public commitment is "this software is designed to run the following: MS Word Lotus Notes Quicken" and it actually works, you have a believer. You have someone who will get excited and will offer their help. You give momentum to Wine. You get program sponsors (those willing to oversee ongoing supervision/testing of programs. You get more people offering to be on documentation teams. And you get more developers.
This is well understood by most people, and this is what we call 09/1.0. That's why it is so important to nail the tasks left for 0.9. You see, 0.9 is not 'all chords', but rather 'a handful of notes' such that we start 'playing a few tunes well'. -- Dimi.
At 10:47 PM 3/19/02 -0800, Michael Robertson wrote:
The challenge we have now is that the goal for WINE is an architectual one (learn every chord). While nobody can deny the value of a solid underpinning, users don't care about the architecture. They want to hear music! And with Wine this means "what programs can I run?"
I agree!
You get more people offering to be on documentation teams. And you get more developers.
Right! Very good article. I just want to add one thing: We need a centralized site where users can vote what programs they want supported on WINE. I think Transgaming has already done this for their subscriber base. I think that codeweavers has something similar. I would be good if we could put that databases together. Maybe we should also categorize a little, like TOP-10 games, apps, etc... Roland
--- Roland <roland(a)netquant.com.br> wrote:
We need a centralized site where users can vote what programs they want supported on WINE. I think Transgaming has already done this for their subscriber base. I think that codeweavers has something similar. I would be good if we could put that databases together. Maybe we should also categorize a little, like TOP-10 games, apps, etc...
I suggest to use Codeweavers Wine application database (http://wine.codeweavers.com/appdb/). I think it had votes information before, but I can't find this information now. If Transgaming have the database it is more XWine-oriented. We, probably can exchange information between these 2 databases, but I doubt we'll be able to completely consolidate them. Andriy Palamarchuk __________________________________________________ Do You Yahoo!? Yahoo! Sports - live college hoops coverage http://sports.yahoo.com/
On Wed, 20 Mar 2002, Andriy Palamarchuk wrote:
--- Roland <roland(a)netquant.com.br> wrote:
We need a centralized site where users can vote what programs they want supported on WINE. I think Transgaming has already done this for their subscriber base. I think that codeweavers has something similar. I would be good if we could put that databases together. Maybe we should also categorize a little, like TOP-10 games, apps, etc...
I suggest to use Codeweavers Wine application database (http://wine.codeweavers.com/appdb/). I think it had votes information before, but I can't find this information now.
Yes, the application database lets people vote for the applications they want working on Wine most. * You can see the applications sorted by votes: http://appdb.codeweavers.com/votestats.php * And the instructions about how to vote are here. You have a total of 3 votes. http://appdb.codeweavers.com/help/?topic=voting What seems to be missing is a count of the votes for a specific application on the application's page itself. You can also vote for the bugs you want to see fixed most. * To vote for a bug you want to see fixed, go to that bug and then click on 'Vote for this bug'. This sends you to another page. Set the number of votes you want to spend on that bug (you are allowed a total of 5 votes), and then click on 'Submit'. * You can view how many people have voted for a given bug directly on that bug's page. * Use the following query to get a sorted list of the bugs by number of votes: http://wine.codeweavers.com/bugzilla/buglist.cgi?bug_status=NEW&bug_status=A... Of course few bugs and applications have votes. For instance only 6 applications have more than 3 votes (if >3 votes then there is more than one person interested in the application). So I would encourage everyone to go and vote for their favorite application. I believe this is the best way to build a Top 10 list. Btw, the current top 10 is: 1. Half-Life 40 2. EverQuest 7 3. Microsoft Word 6 4. Diablo II 5 5. Dreamweaver 4 6. Quicken 4 7. MechWarrior 4: Vengeance 3 8. Director 3 9. Max Payne 3 10. Internet Explorer 3 -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ Linux: It is now safe to turn on your computer.
I think millions of people have voted already on the titles they'd like to run and the results are evident by actual consumer usage of software titles. The list that Steven made seemed solid to me. -- MR
How about this list? http://www.amazon.com/exec/obidos/tg/browse/-/294964/ref=sw_h_r_b_b/103-2241... or this one http://shopper.cnet.com/shopping/0-7076-310-0-0.html?tag=pop Michael Robertson wrote:
I think millions of people have voted already on the titles they'd like to run and the results are evident by actual consumer usage of software titles.
The list that Steven made seemed solid to me.
-- MR
participants (11)
-
Andriy Palamarchuk -
Bernhard Rosenkraenzer -
Dimitrie O. Paun -
Dmitry Timoshkov -
Francois Gouget -
Goran Thyni -
James Hatheway -
Michael Cardenas -
Michael Robertson -
Roland -
Steven Edwards