|This is the talk page for discussing improvements to the WinFS article.|
|WinFS has been listed as one of the Engineering and technology good articles under the good article criteria. If you can improve it further, please do so. If it no longer meets these criteria, you can reassess it.|
|WikiProject Computing / Software||(Rated GA-class, Mid-importance)|
|WikiProject Microsoft / Windows||(Rated GA-class, Mid-importance)|
- 1 Where the article fails searchers: WHAT FOREST?
- 2 Other impls
- 3 Page move?
- 4 Rewrite
- 5 "Architecture" section
- 6 Near complete rewrite
- 7 Backporting
- 8 Desktop Search?
- 9 Criticism of necessity
- 10 Windows to Drop WinFS
- 11 Motivation
- 12 ReiserFS
- 13 "Failed Microsoft initiatives"?
- 14 WinFS as Data Storage backend
- 15 GA Notes (failed)
- 16 Only works with NTFS?
- 17 pre GA review comments
- 18 GA Review
- 19 WinFS does NOT work on top of FAT32
- 20 Merge
- 21 protogon. refs
- 22 Project Start
Where the article fails searchers: WHAT FOREST?
-- there are all these TREES in the way! The article should be structured with less emphasis on MS esoterica and more on practical ALTERNATIVES that are in the marketplace (if Microsoft hasn't driven EVERY LAST ONE OUT of the marketplace by using this "just wait till" tactic for the past 12 years) in order to accomplish the basic KM workflow that the MS people have implemented in very limited fashion via VISTA DETAIL (e.g. tagging, and reviewing) ONLY for office docs and JPEGS.
To me you look at this history and you see that MS is constitutionally NOT CAPABLE of creating an open-system filing system without destroying some sense of proprietary control. It's that simple. Therefore they HAVE TO HAVE demented, inpenetrable technologies, so that they can retain control and simply say "well, you don't understand." This article as it is, is a poster child for MS dementia. If I wanted to be insane I would drop acid and read this and actually try to understand it -- much less as a developer to worry about mastering this kind of pure crap. TELL US WHAT WORKS. How do we get from what is solid (the mechanical file system) ACROSS THE RIVER STICS HERE to the shore of Okay, now I can do this complicated job of mine. Heh?
FOLLOW UP: (Told you so!) For the record Win7' s Library system is a snazzy implementation of the technology that would be very easy to back-implement into a universal advanced file system. For that matter the original SHORTCUTS technology implements just such a feature by linking the file system U.I. to the file index -- something apparently no 2nd party developer has direct access to, making self-healing links native to our daily applications 'difficult.' Why am I not surprised that after 12 years of this BS we still do not have a simple piggy-back Db system that could be local-file based (like the Win7 Libraries) to give us the meta data that can be saved in-file on MP3, JPG, and Office files? Having looked I can find only semi-obsolete developments of a "file note" technology -- the most modern of which is in PowerDesk Pro 6 & 7 -- another being "FNO" - File Notes Organizer that is apparently abandonware. PD however still works well, and has enough of the Advanced File System feature to get me by on my very demanding UUNIS.org project. The page of notes I can take is more than sufficient and I can just use ***** as a pattern in the front of the thing and you can have a rating system that isn't as HOKEY as the one in Explorer that takes up three times the space -- and you can change the font -- get a popup look instantly at the entire first line of text that can be quite long -- very cool. Integrates directly into details view and on the properties / details tab. Move the file with PD and you keep the notes intact. I don't know if there's a hidden file or a small Db being used. And in spite of the serious work potential, PD (Avanquest) doesn't even know it exists or how to sell it. DORKS. There's exactly ONE LINE about it anywhere -- and (pay attention) IT ISN'T IMPLEMENTED BY DEFAULT (add "Notes" to the detail view the usual way.)
I probably just sold ten thousand five hundred packages by mentioning this -- hey fella's -- where's my check? Damn it doesn't work that way. When I figure out how it does work I MIGHT back-engineer the previous paragraph to let you in on it.
- If you are able to have a civilized conversation in sane English about how to improve the article, please post. If all you want is rant, go elsewhere. WP Talk pages aren't designed for that. And btw, if it were so easy to implement, why is it that no general-purpose file system does so? And FYI, if metadata is all you care about, Vista onwards, the entire platform is in place. Just plug in your filetype-specific metadata reader/writer and you are good to go. --soumtalk 03:41, 25 June 2009 (UTC)
- Not really. BFS supported metadata via alternate streams and had an indexer as a part of the file system. So it supported querying the file system, and those could be more capable that a grep-like search. That was it. There wasnt much support for relationships as a first class citizen. From that perspective, these things (BFS, WinFS, Reiser4 etc) could be considered of the same lineage. --soum talk 05:38, 17 October 2008 (UTC)
I have added a link to an old F/OSS project which implemented a similar system on top of GNU/Linux to the "External Links" section. Edit as you see fit. 188.8.131.52 23:48, 19 September 2006 (UTC)
Originally scheduled for longhorn? Come on! It was promised at least as early as 1994, well before Microsoft had any concept of Longhorn. Anyone have any reference from that long ago, I'd like to update this page. --Yamla 00:04, 2004 Dec 15 (UTC)
- Yes, WinFS was originally scheduled for longhorn. What was promised earlier (for Cairo), was called OFS, not Longhorn. And that is properly covered in dev section. --soum talk 15:05, 30 June 2007 (UTC)
Should we link to similar technologies such as Reiser4?
- Does Reiser4 have anything to do with database concepts, metadata? I couldn't tell from the existing article. i.e. how are they similar technologies? Jeff schiller 20:01, 17 October 2005 (UTC)
Done, made the changes as discussed. May come back and rewrite slightly. --Yamla 15:47, 2004 Dec 30 (UTC)
I question if vaporware really merits an article. Does anyone have any good whitepapers to provide some scientific background to give this article some value? Gmaxwell 23:28, 27 Dec 2004 (UTC)
- Don't the links in the reference section provide you with enough scientific background? --soumসৌমোyasch 14:01, 5 February 2007 (UTC)
- I'd say wait. Microsoft will probably change their minds a few times before the thing is actually released. WinFS is better for now, because that is what most people use to refer to it. AlistairMcMillan 15:55, 9 September 2005 (UTC)
I rewrote pretty much everything. Notably: Less speculative, more details; a section on "why WinFS?"; revised release schedule timeline; links to OS X components that are comparable to parts of WinFS; information about prior PDC release. In time, it'd be nice to see this article gain the kind of detailed explanation that similar API articles (e.g. MDAC) have. Warrens 19:17, 17 December 2005 (UTC)
The first paragraph of Architecture needs some help. Some of it is unclear (e.g., "like the thing we names as file system") and it shifts to first person. I've copied it here but haven't modified the original.
"WinFS is built on top of NTFS which provides journaling. WinFS works like the thing we names as file system. We can place standard files into it, but also WinFS is the place where we can store information that is not placed in files but is a representation of the related tables. Yes, WinFS is built from Yukon code. Most parts of WinFS are written in C++. The set of managed APIs are surely written on C#, and as APIs WinFS Services are also written in C# such as WinFS Sync, WinFS Rules. But core parts are written in C++ as they interact on down-level with Windows OS Kernel to get an access to memory and so on."
Near complete rewrite
The article was not comprehensive enough. It felt like some random notes taken in a presentation. Nor was all of WinFS' capabilities covered. I rewrote it exhaustively. Please comment.
I just rewrote the long article. My fingers are numb with exhaustion. Please help wikify.--Soumyasch 04:04, 4 April 2006 (UTC)
- If someone feels the new version is not good enough, let's have a discussion before it is undone. --Soumyasch 04:05, 4 April 2006 (UTC)
- Wow, nice one, Soumaysch! WinFS There is some refinement and flow work to be done, but having the information in there to work with gives us a good starting point. I'm really not a fan of advertising words like "rich" to describe API's, so I may go in and neutralise some of the descriptive language a bit over the next few days. It's one of my favourite subjects... I kinda got this ball rolling back in December, because the subject definitely deserved a good article and I was getting tired of explaining the real WinFS to people. :-) Warrens 06:05, 4 April 2006 (UTC)
- Thanks Warrens. The article now just needs to be refined. --Soumyasch 08:36, 4 April 2006 (UTC)
I oppose WinFS being categorized as a Desktop-Search environment. It goes much beyond that. I agree, initially the most prominent use will be for searching, but its usage is not limited to that. It is a full blown relational database, plus includes a replicator. A lot many things can be done. I myself made a p2p search engine (using WinFS, MS Rave and some others), which, despite not having any centralized index, searched info from all over the network. So, how can it be called just a Desktop search. --soUmyaSch 05:14, 25 May 2006 (UTC)
Criticism of necessity
This essay, famously includes the following passage regarding WinFS:
But Microsoft needs to give you a reason to buy Longhorn, and what they're trying to pull off is a sea change, similar to the sea change that occurred when Windows replaced DOS. The trouble is that Longhorn is not a very big advance over Windows XP; not nearly as big as Windows was over DOS. It probably won't be compelling enough to get people to buy all new computers and applications like they did for Windows. Well, maybe it will, Microsoft certainly needs it to be, but what I've seen so far is not very convincing. A lot of the bets Microsoft made are the wrong ones. For example, WinFS, advertised as a way to make searching work by making the file system be a relational database, ignores the fact that the real way to make searching work is by making searching work. Don't make me type metadata for all my files that I can search using a query language. Just do me a favor and search the damned hard drive, quickly, for the string I typed, using full-text indexes and other technologies that were boring in 1973.
I am not saying I think that this article needs a Criticism subsection, but if there is a good reason to implement this whole database system, rather than just doing what Google does in making search work, the article should explain it and cite this seminal article. —The preceding unsigned comment was added by 184.108.40.206 (talk • contribs) .
- He wrote that article two years ago -- before most of this Wikipedia article could have been written, because the information simply wasn't out there. WinFS is about creating relationships between data, not searching for it. Microsoft didn't do a good job of explaining any of this until mid-2005. Warrens 15:25, 27 May 2006 (UTC)
- The whole section, starting with the title, is blatantly POV. I've fixed the title. I don't have time at the moment to fix the section, i.e. by making the "this is why this is good" stuff into "here is the perceived problem and here is how it was addressed (and probably a pointer or two to other reasonable approaches) -Dmh 15:26, 10 July 2006 (UTC)
Windows to Drop WinFS
see http://blogs.msdn.com/winfs/archive/2006/06/23/644706.aspx for details
I was thinking about changing the "pictures which have person X" example because it can be confused with "face recognition" which I think it's not the feature. Maybe it should be changed to "documents which have X word(s) on them" David Morón 19:01, 18 September 2006 (UTC)
- They aren't the same things though... "pictures which have person X" suggests a relationship between two different types of data. "documents with have X words in them" is not a relationship in this nature; it's just a matter of full-text search... it doesn't demonstrate the concept being discussed. -/- Warren 05:43, 19 September 2006 (UTC)
- Then another example should be used because of the reason stated above, i.e. confusion with face recognition which it's not the feature. Any suggestions? David Morón 21:04, 23 September 2006 (UTC)
- I dont think it is necessary to update the example. Suppose, it is changed to something like "a video which contains sting rays", then also your question is valid. Or something like "a document that refers the 'P' hotel" also raises the question how is it identified. However, according to me, the context makes it pretty clear that WinFS is not about how is it identified - whether pattern recognition is used, or manual tagging, or some other rocket science technique - rather it is about what is done when things have already been identified. Also your suggestion doesnt hold, because WinFS is not interested in which document what words are there. Rather, it is bothered about the semantic structure of the document, i.e., what information does it provide. In that sense, the example is perfectly valid, because "a picture containing a person X" also has this info in a sequence of words only, but that is not essenced by WinFS, rather it deals with the fact that the picture is of person X. --soumসৌমোyasch 11:39, 24 September 2006 (UTC)
- There are some overlap in functionality between WinFS and ReiserFS 4. But ReiserFS 4 is only in development and not mainstream. Reiser3 doesn't have these. A better comparison would be Storage, but that is also in development. --soumসৌমোyasch 10:19, 17 October 2006 (UTC)
"Failed Microsoft initiatives"?
I'd question WinFS's inclusion in the category "Failed Microsoft initiatives". It's still in development & hasn't been released yet, so it can hardly have said to have failed. Delayed, maybe; but that hardly puts it alongside Microsoft Bob. In accordance with Be Bold, I'm removing it from that category. Simxp 03:41, 21 February 2007 (UTC)
WinFS as Data Storage backend
"Microsoft plans to release all of these products [Windows Mail in Windows Vista, Windows Calendar, Windows Photo Gallery, Windows Media Player, and Microsoft Office applications such as Microsoft Office Outlook]with WinFS as the data storage sub-system in the next 3 years"? :O Am I dreaming? --soumসৌমোyasch 08:04, 21 February 2007 (UTC)
GA Notes (failed)
- The motivation section drags on and on. What this section gives you for this article could be summarized in a paragraph. The rest of it would do better in and article named Comparison of WinFS and NTFS.
- The use section states "[Integrated Storage] will be of great use for businesses, by allowing it to automatically aggregate data from different departments" with no citation.
- The development section could use citations.
- The Data Storage section and the Architecture seem too related to deserve different sections. (Most of the stuff in Architecture could be merged to Data Storage.)
- Many of these sections could use more links to standards (eg. XML schema).
- Perhaps take a look at Wikipedia:Explain jargon.
Again this article needs citations. Gutworth 02:03, 16 April 2007 (UTC)
- The article has been streamlined since last nom, references inlined (rather than having them in a generic refs secn), wikilinked and descriptions added and redundancy reduced. Overall, its in a much better shape. I am nominating again. --soum talk 15:02, 30 June 2007 (UTC)
Only works with NTFS?
I keep reading that WinFS runs on top of NTFS. I know this is true, but does it not also work with FAT and FAT32? Josh the Nerd 17:35, 26 June 2007 (UTC) Josh
- No, the only beta release that ever came out worked on top of NTFS. (Even Vista can use only NTFS for its boot volume). --soum talk 18:13, 26 June 2007 (UTC)
pre GA review comments
- Some of those sentences are a bit dense. Where there is a sentence that goes over three lines, it would be a good idea to chop it into smaller pieces.
- The introductory paragraph is a bit dry.
- It would look good to have some kind of picture near the top to grab interest. Or possibly an infobox of some sort.
- related together by Relationships — no capital R.
- The applications all just seem to be support programs. Are there any real applications, or any intention by Microsoft to put a critical operating system function in it - eg the registry?
- One of the references has an error: Error on call to Template:cite web: Parameters url and title must be specified
- An excessive number of sources are blogs - are these reliable?
GB 11:22, 3 July 2007 (UTC)
- Thanks for your comments. I would have to take a look at long sentences. As for the lead, I actually toyed with a lot of variants, but any more description was getting full of jargon.
- As for applications, WinFS is primarily a developer API, not an end user application. As such, all tools released (at least around Beta 1) will be developer tools only. As for, "Are there any real applications, or any intention by Microsoft to put a critical operating system function in it?", well, the plan was to move entire data organization system to WinFS (replacing Explorer and its file management capabilities). But WinFS got cancelled before it could happen in Windows Vista timeframe. Since the technology is still under active development, we have to wait till Microsoft loosens up a bit on the next iteration of Windows before we know what how they are going to integrate the capabilities into Windows.
- As for images at the top of the article, well what can be added? Its a storage subsystem which can store structured data items (think XML and relational data) which are related with each other. How can this be shown in a screenshot. Even if a schematic representation of the association is shown, it will look more like a relational schema). I dont think that can be of interest to users.
- However, I think, screencaps of the IWish concept video, which was shown by MS as user experiences that could be possible with shared data schema and relationships, could be used here. That gives me an idea - in the lead, we can explore this vision to show what WinFS could achieve without jargonning the hell out of readers.
- As for the blogs, yes they are reliable. The blog entries are the official blog entries of the team developing WinFS at Microsoft. Its canonical, at least for technical details. --soum talk 11:48, 3 July 2007 (UTC)
- Thanks for you efforts in replying soum.
- So on the applications front, are there any other companies that have made used of the API or WinFS to do something?
- For a picture on the top - a diagram along the lines of what you mentioned would be good: supplied support tools/ on top of API/ over the top of winFS / over thje top of NTFS(s).
- A timeline of announcements, delivery, or failure to deliver could be included. There is special support for timelines in wikipedia.
- A screen dump of something could be good, as long as you dont pinch it from the microsoft video!
GB 12:03, 3 July 2007 (UTC)
- The API was still incomplete when WinFS, at least in its last known incarnation, was scrapped. As such, third party support was limited. As such, most technologies using WinFS moved to other technologies (as an example, Internet Explorer 7 in Windows Vista was supposed to use WinFS to store its history and bookmarks and include a media player like organization capability. But due to lack of WinFS, that feature got cut out. Also MS Office 2007 was supposed to use WinFS, but they moved to a different platform. And no, I do not have source for this, this is based on leaked info and stuff. Thats why it is not in the article :) ). Because the applications were also half-baked, its best they are left out.
- A diagram of the architecture stack may be included. And the timeline sounds a damn good idea. A screenshot of the StoreSpy application will probably be a good idea. But the main problem is that the preview release they made available was time-limited and my copy has already expired. They are not allowing further downloads of it either. So, it has to be taken off the net. I think this makes a very good Fair Use justification - it is not available publicly, it has been cancelled, and it is a solution that works at a very abstract level and no other market alternatives exixts. So, a picture is very essential as to understand how interaction with the system is done. The same rationale applies for screenshots of the concept video. The video played an important part in building the hype around WinFS and it was used extensively by MS for demonstrations. So, I think it can and should be used. I will add text regarding the use of the video and the concepts in it. --soum talk 12:29, 3 July 2007 (UTC)
- The pictures and diagrams look good with good fair use rationale! GB 21:48, 4 July 2007 (UTC)
- It is reasonably well written.
- a (prose): b (MoS):
- It is factually accurate and verifiable.
- a (references): b (citations to reliable sources): c (OR):
- It is broad in its coverage.
- a (major aspects): b (focused):
- It follows the neutral point of view policy.
- a (fair representation): b (all significant views):
- It is stable.
- It contains images, where possible, to illustrate the topic.
- a (tagged and captioned): b (lack of images does not in itself exclude GA): c (non-free images have fair use rationales):
- a Pass/Fail:
PGWG 16:47, 9 July 2007 (UTC)
WinFS does NOT work on top of FAT32
The article states that WinFS is implemented as a database on top of FAT32 and cites a reference. I thought it was pretty odd that they would use FAT32 instead of NTFS, so I looked at the reference, and the term "FAT32" doesn't occur anywhere in it. It does, however, say that WinFS is built atop NTFS. Someone earlier in this discussion page also mentioned that the only beta release used NTFS. Inhahe (talk) 04:27, 30 June 2008 (UTC)
- Oppose strongly. The two are not even remotely related. Just because they have *FS in their name, people think they should be merged? This is nothing to merge - xpclient Talk 18:40, 17 January 2012 (UTC)
- Agreed. It's not only uncertain, but complete speculation. - Josh (talk | contribs) 20:45, 4 January 2012 (UTC)
What's to merge? ReFS is mentioned briefly in the Win8 article - essentially nothing (reliable and verifiable) is known about ReFS so far. There's nothing to merge (yet) - Meewam (talk) 06:09, 15 January 2012 (UTC)
searching it will reveal that win8 pre-release builds had a file system that seems to be a continuation of winfs. protogon was it's name and then refs. don't have time at time of writing but they should be in this article. — Preceding unsigned comment added by 220.127.116.11 (talk) 20:13, 21 March 2013 (UTC)
- Despite some early speculation that Protogon/ReFS was a reincarnation of WinFS, the two are unrelated. - Josh (talk | contribs) 22:58, 21 March 2013 (UTC)
About 1994 or 1995 Microsoft (in Vienna, Austria, Europe) offered a web-survey with just a very small text-field for general propositions. So I sent a brief concept to them, which covered the principles of an object-orientated (non-redundant) information management, as described in this article. I had been starting developing it in 1988 - the first version was a simple non-restricted database with a suggest-search-field for item-names, similiar to nowadays Wikipedia (but without social functions, as there was no public internet at that time). I never received a feedback and wondered if the messages had been evaluated by departments who understood them and thought about or if the survey mainly covered people's opinion about existing software or maybe the idea was seized and adapted to Microsoft concepts. Does anyone know about the origin of the project within Microsoft? -- 18.104.22.168 (talk) 07:43, 12 April 2014 (UTC)