Spreadsheet formula bugs

아티클
07/12/2007

I saw this blog post from Rob Wier the other day, and I thought it would be worth drilling into a bit more. While Rob tends to have a flare for the dramatics he does point out a couple of good bugs in the formulas piece of the Open XML spec. Rob's actually been spending a ton of time lately pouring through the spec looking for issues and representing those issue on a number of different national standards bodies. I'd like to give Rob the benefit of the doubt and assume that he's doing this in order to help improve the spec, and not just to try and find a reason to block it. It would have been much more efficient if he would have worked with us on TC45 though when we were initially reviewing and working through the spec. I know everyone on TC45 would have appreciated having him aboard, and since IBM is already an Ecma member it would have been simple for him to join. Either way though I appreciate the time that Rob is spending on the spec, as it's something you all know I care deeply about.

Let's talk about the bugs he pointed out in the formulas spec. It's a rather long post, but can be filtered down into the following 7 issues:

The type of units on the values value for trig functions are not specified (radians or degrees)
AVEDEV function is using an incorrect formula
CONFIDENCE function needs some more information
CONVERT doesn't specify which "cup" or "tablespoon" version to use
Day count basis settings such as "US (NASD) 30/360" don't actually point to a full definition
NETWORKDAYS function only assumes one style of a workweek
Some statistic functions say "x is the sample mean" rather than "x-bar is the sample mean"

I took a look at the spec and it looks like he's right, there should be a bit more information provided and errata corrected. That's something that should be fairly easily handled during the ballot resolution meeting. The way the fast track process works is that every country gets a chance to vote on the spec and provide comments. Some groups will vote "yes", but leave comments they'd like to see addressed. It doesn't make sense to say "no" if you just have minor things you'd like to see addressed, which is why the ability to leave comments with a yes vote can be useful. Other groups will vote "no", and provide a list comments that explain what led them to that no vote. At the ballot resolution meeting, we get a chance to go through all the comments and make any necessary changes to the spec.

Some comments will probably get dealt with at the ballot resolution meeting, while others may be taken into account for future versions of the spec (depending on what everyone decides). In ODF for example, accessibility was probably raised as an issue, but it still was approved by ISO. Then OASIS went off to work on version 1.1 of ODF and added improvements specific for accessibility. Now they can take the 1.1 version to ISO and seek to get approval for that. The continual improvement and evolution of the standard is just as important as the initial review and approval. It will be important for Ecma and the ISO to define how the Open XML spec will evolve in the future.

In regards to the comments that Rob made around spreadsheet functions, I would hope those could pretty easily get handled at the ballot resolution meeting by just making the following corrections:

#1 The type of units on the values value for trig functions are not specified (radians or degrees)

This is easy enough to fix. The measurements are all in radians, so that can be easily clarified in the spec.

#2 AVEDEV function is using an incorrect formula

Looks like the wrong formula was inserted. I almost wonder if this was just a copy/paste screwup… It should be this:

#3 CONFIDENCE function needs some more information

Looks like we should have made it clear that this assumed a Normal distribution.

#4 CONVERT doesn't specify which "cup" or "tablespoon" version to use

If this wasn't clear then we should probably provide a link to the NIST definitions: https://physics.nist.gov/Pubs/SP811/appenB9.html#VOLUME

#5 Day count basis settings such as "US (NASD) 30/360" don't actually point to a full definition

A number of other applications do the same thing (including the OpenFormula spec Rob is working on), so I thought this was fairly widely known by folks working in this area. Given Rob's feedback though it probably would have been better to have included a reference to this:

Standard Securities Calculation Methods: Fixed Income Securities Formulas for Analytic Measures, Vol. 2, Spring 1995.

#6 NETWORKDAYS function only assumes one style of a workweek

Yeah, this is another one of those issues left over from the Lotus 1-2-3 days. The function has always behaved like this, and we never changed it. It could break existing solutions if we changed it now, but it would also be a good thing to add a new function that mimicked the old but also had an additional parameter to specify the other work week types. Not sure if the creation of a more extensible function is better to handle in the ballot resolution meeting or in the next version of the spec. I guess we'll see.

#7 Some statistic functions say "x is the sample mean" rather than "x-bar is the sample mean"

Yeah, looks like another bug. Rob's right that it should say "x-bar" instead of "x"

So, as I'm sure anyone would expect with a specification this large there are some errata that need to be addressed. I can't imagine any spec could be bug free, and thankfully we have a lot of people interested in this space who will help us as we continue to improve on the spec going forward.

I did want to get one little jab in at Rob though around his implication that we rushed thing with Open XML and that ODF made the right choice in delaying the formula stuff. In reality the OpenFormula technical committee didn't even exist until a year after the ODF spec was finalized and a few months after Microsoft submitted Open XML to Ecma. The people working on the group are a completely different set of folks from the initial ODF creators. The main folks who worked on ODF (ie participated in at least half of the meetings) were Gary Edwards; Michael Brauer (Sun); David Faure (KDE); Doug Alberg (Boeing); Daniel Vogelheim (Sun); Patrick Durusau; and Paul Grosso (ArborText). The OpenFormula committee is primarily made up of Rob from IBM, Eike Rathke from Sun; and David Wheeler from the OpenDocument Foundation (Jody Goldberg who worked with us on OpenXML also participated for awhile). So I think it's a bit disingenuous to make it sound like they already had a formula spec underway but decided to hold it back because it wasn't yet ready for prime time. :-)

-Brian

Comments

Anonymous
July 12, 2007
Could you please comment on what I think is a very pertinent post for OOXML issues (it seeks to explain why OOXML is of such importance to Microsoft): http://www.microsoft-watch.com/content/business_applications/obla_de_oba_da.html ?
Anonymous
July 13, 2007
Hi Brian, Since we're on the topic of vague documentation, I have a question concerning the colour tint. Now aside from the fact that the tint is stored/applied in different ways depending on whether it is Word, Excel, etc, it isn't a problem for me. Where I'm having trouble is in the DrawingML version of tint. Actually, I might as well link directly to my query in the OpenXML Developer forum http://openxmldeveloper.org/forums/thread/1627.aspx The problem with getting the tint wrong is that it makes it really difficult to know if you are implementing the already complex chart styles correctly because one uses tint to alter colours all over the place. Unfortunately, as Stephane pointed out in a previous comment, the Forum can at times be pretty unlively. My query hasn't been responded to in a month and a half. I realize everyone is working on the next release of Office, but could a couple MS employees be assigned to answering questions from time to time? Maybe an intern can run around finding answers? I've tried answering people's questions before, but the moment I reply, no one looks at that post again, so if I'm wrong I'll have misled the poor person. Thus I refrain from replying. And if no one answers, no one will bother asking. And then the forum really will die. BTW, thanks for the post on the Word document model. It finally answered the great mystery of why section properties are stored at the end of the section. We worked around it, but it was nice knowing why we had to :)
Anonymous
July 13, 2007
The comment has been removed
Anonymous
July 13, 2007
Daniel, I think it's pretty clear if you read that article that Joe doesn't really understand XML. He keeps talking about proprietary schemas and a dependency on XSD. We use XSD to help provide validation information for our XML, but the file format isn't dependent on XSD (we also include RelaxNG for folks who'd rather use that). He also says that translation is necessary only because we use "proprietary schemas". Moving from one format into another always requires translation. I've talked to Joe before about this, and he mistakenly believes that once everything is XML no translation should be necessary. That's obviously not true, translation will always be needed to go from one format into another. By using XML it just gets much easier. Andrea, I'll go talk to some of the DrawingML folks and see if they can post a reply to your answer. I'll also see if that's something we should recommend to Ecma to put more info into the spec. hAl, I just hate the thought of someone having to spend their time trying to halt progress, not help it. I can't imagine any engineer getting satisfaction out of that. I don't care how much you get paid, life is too short. We've definitely been talking within Ecma about how to deal with all of these errata, and requests for more information. I agree with you about the depth of this review, it's incredible. To your last point, it's hard for Microsoft to commit to what comes out of Ecma in the coming years, because we don't know what direction they will take the formats. We'll of course stay active and propose changes based on where we want to go with Office 14. At the end of the day though, the other Ecma members could decide to take the spec in a completely different direction. Now my impression is that won't happen, as the folks on the TC all have pretty similar visions for the future of the spec, but since it's not guaranteed it would be hard for us to make any sort of official statement. In terms of licensing, we can't provide licenses for new stuff Ecma adds, since Ecma owns it. Our license applies to everything we've submitted, and if we submit anything new we would probably just use the same license. What kind of statement do you think would be useful (given the fact that we can't predict the future)? -Brian
Anonymous
July 13, 2007
I've recently discovered this site, where people are obviously going through the tiniest details of the documentation, right down to typos, copy/paste errors, etc. MS should take advantages of all us developers who are staring at the documentation and finding all these tiny errors for them. Combined with the Forum you're getting free copy editing and QA of the format! http://www.xmlopen.org/ooxml-wiki/index.php/DIS_29500_Comments
Anonymous
July 13, 2007
Ecma will definitely take advantage of it. I wish everyone had looked this closely at the public drafts we were releasing during the development period last year. -Brian
Anonymous
July 13, 2007
more errata: s/flare/flair/ s/dramatics/dramatic/ s/pouring/poring/
Anonymous
July 13, 2007
[quote]What kind of statement do you think would be useful (given the fact that we can't predict the future)?[/quote] I guess it should be about committing to further development of the ooxml standard together with Ecma and as you just stated to commit for keeping open any licensing for the features in the standard that Microsoft contributed to the standard or will contribute in future.
Anonymous
July 13, 2007
NitPickerA, You should pass those on to the noooxml.org site and see if you can win the 2,500 euro prize :-)

hAl, I'll dig around as I'm almost positive we already have a statement saying those exact things. -Brian

Anonymous
July 13, 2007
The comment has been removed
Anonymous
July 13, 2007
I think that OpenXML is a huge progressive step forward from the old world of the binary formats. You don't think so? I believe the world is a better place now that we've moved to using Open XML as the default format. I don't know where Microsoft is going to be in 100 years. I certainly won't be working for them. But I also know that if ISO approves the format, then any files I save from now on will be in a format that is documented and owned by the ISO, so my grandchildren, and their grandchildren will be able to read what I've written. What good does it do trying to block it? Why spend a bunch of time and money on blocking it unless you feel like you're going to use the fact that it's not ISO as a competitive advantage for you? Open XML files are going to exist. Over the coming years there will be hundreds of thousands of open xml files created. Only a fool would disagree with that. So, given that this will happen, wouldn't it be better if the specification for those formats was owned and maintained by the ISO? -Brian
Anonymous
July 13, 2007
Brian You said it. In yr reply above simply replace OOXML with ODF and what u get is what all the proponents of ODF have been shouting hoarse about.
Anonymous
July 13, 2007
The comment has been removed
Anonymous
July 14, 2007
The comment has been removed
Anonymous
July 14, 2007
The comment has been removed
Anonymous
July 14, 2007
I think there's a happy medium between "political" and "technical" stories, because many of the political arguments made on this blog are technical arguments at heart. For example, Microsoft seems to believe that a file format should resemble an implementation as closely as possible, which has drawn scorn from people that believe a file format should represent an idealised view of a document. Although there's been some discussion of this, it tends to produce more heat than light because the examples discussed (although a fine starting point for a debate) aren't traced back to the root issue. That said, you only ever see the deep philosophical underpinnings of your ideas when contrasting them with ideas that have grown up in a completely different environment. So while some of the ODF chatter is a sad necessity, some of it is an excellent way of learning about OOXML.
- Andrew
Anonymous
July 14, 2007
The comment has been removed
Anonymous
July 14, 2007
[blockquote]Yet those same people pretend that ODF is an "idealised view of a document" when it's based on OO.o 1.0's XML format and therefore on OO.o's code structures. It's no more "ideal" than any other format.[/blockquote] I actually think ODF looks more clean and organised as a format spec then OOXML. I actually like that about the spec. However to me the reuse of SVG and MathML quite weak. People name this as a big advantage of ODF but actually I think SVG and MathML are particullary not ment for office document use but for a lot more static webpages with mostly presentational use and not so much editting and that shows. The strong points of ODF as an Office spec cannot be used when using MathML and SVG as elements within such documents. I would have preferred that ODF would have reused those standards but added them in such a way that the elements could fully intergrate with the rest of the xml markup even if that would have ment that it would only make them compatible with those standards. What also scares me a bit is that noone has managed to implement all of ODF yet. Wasn't ODF ment to be a much compacter spec and easier to implement than OOXML ?? There is no prove of that so far. I see Stephan complaining about the lack of information about implementing OOXML but this problem actually seems a lot harder when using ODF as for a lot of features still there is difference in implementations or no example implementation at all. OOXML isn't really implemented much better (mayby even worse in Office 2007) but a lot more is implemented and the spec provides a lot of exmaples. And because Office is marketleader for people implementing there is a fairly complete reference implementation and even when that is sometimes faulty or incomplete on the format spec following it's example will create compatiblity with most documents which is something that organisation require and prefer. Implementations of OOXML will be sellable when compatible with MS Office and not so much when following OOXML better than MS Office does. Therefore it will be very important that Ecma improve the spec and that Micrsoft improves it's support of the spec in MS Office. This is a lot more likely when the spec is an ISO format and Microsoft wants to claim support for that standard. Definitly I am sure that Micrsoft would not want companies claiming that their Office suite has better support for OOXML as an ISO standard than they themselves have.
Anonymous
July 15, 2007
Alon, To say that comments on this blog are too political when you are yourself willing to make sure that everyone sees you are from Israel, is hypocrisy. I'll explain in another comment why Brian spent some of his time to post details about the legacy Word/Excel stacks.
Anonymous
July 15, 2007
Bruno, Ad hominem attacks without putting your real name. How old are you?
Anonymous
July 15, 2007
The comment has been removed
Anonymous
July 15, 2007
Although attacking someone's behaviour and motivations tends to cause more offence than it's worth, it's not really fair to classify an argument as "ad hominem" when it's made in sufficient detail to be susceptible to logical analysis. Bruno's arguments can be confirmed or denied by addressing questions like "is there a hypocrisy in claiming ODF is implementation-neutral?" or "is opposing OOXML's existence an inherently dishonourable tactic?". That said, to my mind these are purely political questions, because knowing the answer tells you nothing about how to write good code, or even about what the ISO should do with OOXML.
- Andrew
Anonymous
July 15, 2007
We are spending way too much time trying to find issues with ooxml. Whose task is it to ensure that the standard is valid? ECMA or ISO? Why is ISO allowing ECMA to put forth this broken document as a potential standard? Why are all of the issues that have been highlighted during the contradictory period and now not have been found at the ECMA level and taken care of? I say it has to do with MS strongarming the ECMA committees into approving the standard at their level without so much as a discussion. To the best of my knowledge, there are no public documents or minutes that explain what happened at the ECMA. I think ECMA has been done a disservice in diminishing it's credibility and standing with this fast tracking of ooxml. I think ECMA should just withdraw this and do a full review.
Anonymous
July 15, 2007
Andrew, Even a cursory of Bruno's comment leads you into thinking this is ad hominem attack. It's actually far worse than that, and I originally did not want to delve into that. But now that you are asking for it, here we go. What Bruno should be doing is thank Rob Weir who is one of the only guys who has spent so much time doing a serious work that Microsoft should have done in the first place, and that Bruno, me, you and others should have done. Brian makes it very clear that what Rob says is correct. Let me say it differently, to call out Rob for his superb work is just a covert to avoid saying that Bruno, me, you and others are too lazy to do it. Which is true. And then, the logical conclusion would be to thank Rob.
Anonymous
July 15, 2007
Nathan: Fine, then let them have their ECMA standard. But why ISO standardization when there is a better format?
Anonymous
July 15, 2007
The comment has been removed
Anonymous
July 15, 2007
"We're all in agreement about that" Re-read Bruno's comment. "Certainly Rob Weir has done a lot of valuable work tracking down bugs in the OOXML spec." It's not even a spec. Everyone calls it a spec because someone did first and then everybody else followed. But this is not a spec at all. It's a description. A spec for the 3 major MS Office formats is somewhere 600,000 pages (and that excludes everything that has nothing to do with it such as user interface, addins and so on). It's a description. Open the ECMA 376 PDF document, pick any page of your choice, and you'll realize that attributes are described with short sentences. I can assure one thing, program managers in the Office team don't specify features this way.
Anonymous
July 15, 2007
I'm afraid I've completely lost the thread of this argument now. In one sentence, what conclusion are you trying to convince me of?
- Andrew
Anonymous
July 15, 2007
In my last comment, I began saying to re-read Bruno's comment since I think you haven't read it right. Again, someone outside Microsoft did a lot of work. There is no way he can be possibly be dissed for doing it, especially from someone (Bruno) who does not seem to have done much (nor do I). In the other part of the comment, I was reflecting on the word "specification", that we keep using. I think it's being abused. And that ECMA 376 is just a description, not a specification. This is concise way to say that this stuff does not deserve to be pushed neither to ECMA (too late I am afraid), nor to ISO.
Anonymous
July 15, 2007
Right, I understand now, thanks. I agree with you that Bruno's behaviour before was disrespectful. My complaint was that an ad hominem attack has a specific technical meaning, and that his behaviour, while rude to Rob Weir, wasn't heinous enough to meet those criteria. As to definitions of terms, I think I see what you mean by "specification" and "definition" - that a specification unambiguously specifies all behaviour relevant to features that are discussed, while a description merely lists those features. As you say, I've not done my homework on this one, but for now I'm willing to take it on trust that ECMA 376 is more like a description than a specification. That said, I'm unaware of any widely agreed definitions for these terms - I suspect everyone here could come up with a completely different, completely sensible definition for either word. While I respect your definitions, and knowing them helps me to understand your position, I hope you won't blame me if I stick to using the words that cause the least confusion when talking to other people.
- Andrew
Anonymous
July 15, 2007
Brian, "I wish everyone had looked this closely at the public drafts we were releasing during the development period last year." Personally, I wished you and other people at ECMA had looked this closely before accepting it as an ECMA standard and releasing it to ISO ! You pretend to ignore that the point made by Rob is the poor technical and editorial quality of the ECMA 376 document (by claiming these are small errors very easy to correct). The sad reality is that MS rushed to try having a document (any document ?) approved by ISO, much more than trying to provide a good document. The hundreds of errors contained in the document, be it serious flaws or easy to correct typos, show that the fast track process was a bad approach. I would recommend that ECMA makes a full review, as suggested by Sum Yung Thing, and prepares a version 1.1 of the document. And by the way, why submitting it to ISO at all ? Do Microsoft think that ECMA is not a respectable standardisation body ?
Anonymous
July 15, 2007
The comment has been removed
Anonymous
July 15, 2007
The comment has been removed
Anonymous
July 15, 2007
I asume several of the ones you do not list as Microsoft related still votes yes to ooxml as well ?
Anonymous
July 15, 2007
"It would have been much more efficient if he would have worked with us on TC45 though when we were initially reviewing and working through the spec. I know everyone on TC45 would have appreciated having him aboard, and since IBM is already an Ecma member it would have been simple for him to join." Given Microsoft's claim that ODF is not sufficient to represent Microsoft's legacy documents, isn't this statement equally valid: "It would have been much more efficient if Microsoft would have worked with the OASIS OpenDocument TC though when they were initially reviewing and working through the spec. I know everyone on the OASIS OpenDocument TC would have appreciated having them aboard, and since Microsoft is already an OASIS member it would have been simple for them to join."
Anonymous
July 15, 2007
hAL, More info here : http://www.robweir.com/blog/2007/07/ooxml-fails-to-gain-approval-in-us.html And here : http://www.ibiblio.org/bosak/v1mail/200707/
Anonymous
July 16, 2007
The comment has been removed
Anonymous
July 16, 2007
[quote]I know everyone on the OASIS OpenDocument TC would have appreciated having them aboard,[/quote] That seems unlikely. Especially as that TC was referred to as the Open Offfice TC for the first few years of it's existence.
Anonymous
July 16, 2007
This is clearly a political debate, but it's also a debate that's really worth having. Since I doubt there's anything Brian Jones would like to add to the debate, the only benefit of having it on his blog is that everyone we're used to talking to is here. How about we all agree to move discussion about INCITS over to http://www.robweir.com/blog/2007/07/ooxml-fails-to-gain-approval-in-us.html, and leave the people that just want to hear technical stuff in peace?
- Andrew
Anonymous
July 16, 2007
If you want to reply to an earlier comment in this thread over on Rob's blog, I suggest you quote the text you're replying to in your message over there, and leave a short message here saying that you've replied.
- Andrew
Anonymous
July 16, 2007
@Andrew [blockquote]How about we all agree to move discussion about INCITS over to 'Rob Weirs blog' [/blockquote] As Brian Jones does not seem to moderate any serious comments I feel a lot more comfortable debating here where I feel that all serious comments are always shown. I know from last year that that is not always the case on Rob's blog.
Anonymous
July 16, 2007
That's a good point, I hadn't noticed the moderation when I suggested that. I'd add that even if he lets through all non-spam comments, it restricts debate to the hours in which Rob Weir is awake and moderating. Could you suggest another location? Doug Mahugh's blog perhaps?
- Andrew
Anonymous
July 16, 2007
The comment has been removed
Anonymous
July 16, 2007
Typo : Those guys are nowhere to be seen, and yet they have voted in favor of OOXML.
Anonymous
July 16, 2007
Stephane, I'm not sure to what degree your comments about Doug Mahugh were aimed at me, but for what it's worth, I'm not suggesting Doug Mahugh's blog because I think he's somehow a neutral party in this all. On the contrary, I suggested it because his blog is blatantly political, so people can't complain about political discussions on a technical blog. That said, if you know of another unmoderated forum where nobody can legitimately complain about us talking politics, say so and we can all move there instead.
- Andrew
Anonymous
July 16, 2007
The comment has been removed
Anonymous
July 16, 2007
The comment has been removed
Anonymous
July 16, 2007
I think we're getting into another debate about definitions of terms. Since there's a substantive issue to address, I'll try and avoid using the word "political" for now. It seems like two types of people read Brian's blog: those that want to talk about whether OOXML should become a standard, what its strengths and weaknesses are, and so on; and those that need to use OOXML in order to get their job done. It seems like Brian's primary interest is in talking to the people who need to use OOXML to get a job done, even if he writes the odd article about goings on in the world of OOXML. It also seems like it's a common theme from "getting a job done" people that they don't understand and don't care about the rest of the debate. If Brian had written an article about the INCITS vote, or if this debate was likely to uncover something that the "doing a job" people would be interested in, I'd definitely agree we should be talking about this here. But as it is, I don't see what it benefits anyone to discuss these issues here, when we could easily go somewhere more appropriate.
- Andrew
Anonymous
July 16, 2007
hAL, Please read my post. You are causing me to post twice as much as I should, only to repeat myself. Those bribe partners are not advanced implementers. There is a huge gap between advanced implementers, and them. If you let me know what is your involvement in this stuff, whether you implement OOXML at all, I can tell you what I think about it. As you might have realized, if you checked the stuff I do, I expose stuff such as calculations and printing in my products. This is clearly not in the same league than others. Your mileage may vary, but I do my market research, competition research, and I know where I am.
Anonymous
July 16, 2007
The comment has been removed
Anonymous
July 16, 2007
@hal said: "[A] minority led by IBM and Sun has easily managed to block consensus over a viewpoint and has just managed to add enough partners themselves to block any kind of consensus vote..." I would like to clarify to you, @hal, what are the possible votes of the NBs in this 5 month review ballot period [1][2]: i) approve the text as is ( with editorial or other comments, but no technical comments ) ii) disapprove the text for technical reasons, and without possibility to change the decission iii) disapprove the text for technical reasons, but if the technical comments are saved by modifications by the text submitter, the vote will change to "yes" iv) abstain One of the votes at Incits/V1 last friday was: "Roll-Call Vote #3: Motion: Jim/Lynn that the recommendation on the U.S. position on DIS 29500 be “Disapproval with comments”. Result: 10 yes, 15 no, 1 abstention. This fails the majority requirement (simple majority of 26 = 14), and also fails the 2/3 rule (10/25 [ignoring 1 abstention] is 40.00%, which is less than 66.67%)." [3] If the MS crowd ( the 15 members that voted "no" to this motion ) would have voted "yes", DIS 29500 could have been corrected and approved ( making a motion for an option iii) vote ) So, the MS crowd blocked DIS 29500, no the other "crowd". [1] ( JTC1 Directives, pages 112-113 ): http://www.jtc1sc34.org/repository/0828.pdf [2] http://www.ibiblio.org/bosak/v1mail/200707/2007Jul02-080502.eml [3] http://www.ibiblio.org/bosak/v1mail/200707/2007Jul13-175415.eml
Anonymous
July 16, 2007
Stephane, Urging hAl to reply on this forum before we've come to an agreement about where this discussion should go is quite rude. I know you value respect very highly, so please treat me with respect by trying not to making a compromise impossible. Getting back to the point, why are we talking about INCITS here? What do you gain by not moving over to Doug's blog?
- Andrew
Anonymous
July 16, 2007
The comment has been removed
Anonymous
July 16, 2007
[quote]If the MS crowd would have voted "yes", DIS 29500 could have been corrected and approved[/quote] That would depend on the extent of those comments and which of which those comments would need to be resolved before the disapproval vote would be lifted. I have no insight in that. Seeing that very strong opponents of ooxml voted for that option leads me to believe that the comments might already be defined in such a way that any possible resolution to lift the disapproval would be out of the question. But you should probably ask Doug Mahugh on his vote choice on that disapproval with comments.
Anonymous
July 16, 2007
hal said "Seeing that very strong opponents of ooxml voted for that option leads me to believe that the comments might already be defined in such a way that any possible resolution to lift the disapproval would be out of the question." The comments were agreed by consensus in the various meetings. Is my understanding that only this comments should be addressed to change the vote. "But you should probably ask Doug Mahugh on his vote choice on that disapproval with comments. " i don't want PR responses, i prefer sincerity !
Anonymous
July 16, 2007
Stephane, I agree that this is a topic worth discussing, but several of our fellow readers have said they'd rather it wasn't discussed here. Since there are lots of places online that we could talk, it seems like quite an easy request for us to comply with, and choosing not to do so would make us look unreasonable. I take your point about Doug's blog though. It's probably too late to move this particular discussion now, so could you suggest an unmoderated, neutral venue that people could use to discuss this sort of thing in future?
- Andrew
Anonymous
July 16, 2007
hAL said "Several of those micrsoft partners actually implement OOXML at some level or implement MS Office related solution." I overlooked an interesting bit in this comment you made yesterday. Microsoft has a strict definition of what makes an MS Office solution. They in fact maintain a portal for just that, the MS Office marketplace (http://office.microsoft.com/en-us/marketplace/). There are two conditions to meet to be listed there : 1) your product is REQUIRED to work with one or more product of the MS Office suite (read:using your product implies someone is also using a license of one or more products of the MS Office suite) 2) Within six months after a major release of MS Office ships, you have to upgrade your product to support it, or you get delisted. All fine, when MS Office solutions serve customers out there (you could argue that this stuff is essentially promotional products helping Microsoft, and that Microsoft can use this as a tool to leverage a partner ecosystem whenever that's needed, but that's beside the point). But then when you talk about MS Office solutions to try to make a point related to OOXML pledges, that's intriguing. Microsoft made a number of bold claims including "platform independence", this just does not stick with MS Office solutions! You can't be both required to work with MS Office (windows), and pledge platform independence at the same time. So this is one more bogus argument that goes along this looonng string of bogus arguments. At the end of the day, there are really two kinds of implementers out there : those competing against MS Office (example : OpenOffice), those working with or on top of MS Office. If Microsoft wanted to give weight to their ISO standard proposal, then sure enough they would want to make sure to get the first kind of implementers to voice their support for it. But, as you can see from last friday's vote, this is only the other kind they got.
Anonymous
July 17, 2007
@Stephane If you provide MS Office solutions then it would seem logical that you implement those on Windows or OS X. but alternatifly you can also use OOXML for solution that are not MS Office related solitions but that create resultfiles that of course by use of OOXML can be opened in MS Office or another ooxml supporting tool like with Novell's OpenOffice support for OOXML. I think the OOXML use by Mindjet and Datawatch Monarch would qualify as such. Thirdly you can produce standard tooling to manipulate OOXML packages files and/or the data in it such als the tools created by Altova. Fourthly you could create programming tools to create and or manipulate documents and data in OOXML files from other applications. Even though MS Office is the biggest platform for an Office documents that does not nescesarily mean that software companies will all focus on MS Office when it comes to OOXML solutions. Actually because of the focus by Micrsoft already on MS Office functions there is a lot of opportunity for Software developers to move that functionality to other platforms where Microsoft isn't so active and the market is much less crowded. I can see tons of reportwriting functionality on all kind of platforms supporting OOXML exports. Simular for mailing tools, digital archiving tools and all kinds of other software products. ISO standardization for OOXML would virtually guarantee good OOXML support on many platforms including Linux and probalby a lot of mainframes as wel as it is commercially viable to create such support then. Even IBM already planned OOXML compatibility in some of it's software allthough they quickly removed that information as it made them look very hypocritical in their stance on OOXML in the standards committees stating that the spec was not good enough to implement whilst they were actually implementing it already.
Anonymous
July 17, 2007
The comment has been removed
Anonymous
July 17, 2007
[quote]OOXML has to provide ONE innovative thing to justify that it's be pushed to ISO.[/quote] Actually it have to provide any innovative thing. A standard has nothing o do with innovation. But compared to the binary formats it has a clear advantage of being able to clearly distingish between data and style whilst keeping all the features and staying compatible with legacy documents. Sepration between data and style is not innovative on an XML level but a big innovation compared to binary formats. It makes it easy to add data and manipulate data in office files even for the simplest of applications where in binary files this was not so easy. You state that OOXML is there to undo ODF as an ISO standard but at the moment ODF in the market is nothing. That is a narrow view held by ODF supporter. To me and most other people OOXML is there to undo the binary format and make the information in those binary formats available on a much much easier level creating a much more level playing field for devlopers in accessing and correctly interpreting that information on any kind of platform. Atm most people really couldn't care less whether or not OOXML crushes ODF or not (allthough I don't think it will). ODF is in real life virtually non-existing whilst binary MS Office files are the standard. Is there any advantage if Microsoft were to use ODF a their main format now ? I cannot see it atm. I would see them make a ton of undocumented propriety extentions to ODF for legacy features though making interoperability of ODF files a joke. ODF will do fine without Microsoft messing with it. There is no reason for ODF to become a the leading document format in a few weeks. If it is indeed a better format then it's growth is likely to be evolutionary with a firm basis in OSS products and Microsoft competitors.
Anonymous
July 17, 2007
"Actually it have to provide any innovative thing. A standard has nothing o do with innovation." When there are already standards for representing this stuff (text, spreadsheet, presentations), what is the reason to push a newer standard, other than innovation? Something that is either game changing, or dramatically better. "But compared to the binary formats it has a clear advantage of being able to clearly distingish between data and style whilst keeping all the features and staying compatible with legacy documents." This proves you have not implemented OOXML and not implemented any of the binary file formats either. The coupling between formatting and content is the same in OOXML than in binary formats. For no other reason that it was the easiest thing to do, and those guys are extremely lazy. That was also the less risky from a compatibility point of view. Even though, in details, they missed that goal to for a number of reasons. Here are two : 1) the compatibility mode is new code and has plenty of bugs 2) the chart drawing engine has killed a number of existing chart properties. By separation between formatting and content, perhaps you are thinking about the "custom XML" scenario. This scenario is the main feature of Office 2003, in case you missed it...This stuff was stored in separate OLE streams. At best, what you can say is that WordML is now up to par with SpreadsheetML, but nothing that justifies breaking the formats (for users out there, all they see are new files they can't open). "You state that OOXML is there to undo ODF as an ISO standard but at the moment ODF in the market is nothing." Sure, its market is not comparable, but the spirit of proper document formats is alive. Is ODF technically superior? Sure. I had to implement styles a while ago, and I was shocked to realize that the simple implementation I had would 1) work with 19 OpenOffice formats 2) would abstract me away from whether styles where being used in text, spreadsheet and so on. When you have spent years of reverse engineering the Microsoft binary formats, this is too good. I don't get that sense when I implement bits of OOXML in my products. At all. "Is there any advantage if Microsoft were to use ODF a their main format now ? I cannot see it atm." Why would not they join the OASIS and start working towards native suppot of ODF? If they believe in interoperability beyond words, that's the next logical step. Anyway, I think the following things will happen in parallel : 1) ODF undone due to Microsoft monopolies on the marketplace 2) OOXML becomes pivot format, ODF, UOF and so on becomes second-class citizens 3) MS Office dies as corporate people move to simpler, secure, better, modern, more cost-effective alternatives.

다음을 통해 공유