Re: [Yaml-core] About the text block productions

[Brian Ingerson <ingy@tt...>]

On 19/09/02 10:12 +0300, Oren Ben-Kiki wrote:
> > 
> >     ---
> >     # Comment
> >     #Comment
> >     ###Comment
> >           #Comment
> >     foo: #comment
> >     bar:#content: baz#content #comment
> 
> So, you want '#' to start a common if and only if it is preceded by
> white-space/line-break/file-start?
> 
> As opposed to today's rule it must both be preceded by and followed by them?
> 
> > This seems to cause no problems. Putting an artificial 
> > restriction on comments by requiring a space where it's not 
> > ambiguous, is a mistake.
> 
> Ambiguity was never a concern here. The question was expressive power and
> intent. When someone writes #bloop, does he mean a comment (as in your
> example above) or does he mean a token (as, for example, in Smalltalk)?

Well, first of all :bloop and #bloop should be quoted IMHO.

I think we'd be remiss if #bloop were content here:

    ---
    foo: 42
    bar: 44
    #bloop: forty-ooops
    ...


> 
> The current spec assumes the latter. Since '#' comments are meant for human
> consumption, the vast majority of them are written with a following space,
> because that's the readable way of writing them - at least in my experience.

In practice, if I'm commenting something out, I'm gonna forget the space most
times. And in an example like this it would be infuriating:

    ---
    - 42
    #- 43
    - 44

What else could this be?

> This allows us the let #bloop be content (just like :bloop is). I think
> that's best; it gives us more expressive power and it more consistent with
> the way that other indicators (: ,) are interpreted.
> 
> So again I'd rather just leave things as they are.

Let's get some more opinions here.

Cheers, Brian

Re: [Yaml-core] draft of YAML slideshow presentation

[Brian Ingerson <ingy@tt...>]

On 19/09/02 14:27 -0700, Brian Dorsey wrote:
> On Thu, Sep 19, 2002 at 12:23:40PM -0700, Brian Ingerson wrote:
> > On 19/09/02 12:04 -0700, Brian Dorsey wrote:
> > > 
> > > -
> > >     exif:
> > >         ColorSpace: '1'
> > >         ComponentsConfiguration: [ 1,2,3,0 ]
> > 
> > This probably won't parse the way you intend it. 
> > 
> >    ---
> >    # This
> >    - [ 1,2,3,0 ]
> >    # Is the same as
> >    -
> >      - '1,2,3,0'
> >    # not (what you intend)
> >    -
> >      - 1
> >      - 2
> >      - 3
> >      - 0
> > 
> > 
> > BTW, If you are parsing this with one of the public tools, you should let the
> > author know.
> 
> Interesting... It seems to work great. ;) I'm using pyYAML, and the YAML
> I posted was all generated by yaml.dump()

*blushes* I forgot to say *why* it's wrong...

The reason is that a separating comma requires a space after it. So:

    ---
    # This
    - [1,000, 2,000]
    # Equals this
    -
      - 1000
      - 2000
    # Not this
    -
      - 1
      - 000
      - 2
      - 000

So, Steve. What does PyYaml do?

Cheers, Brian

PS When I find myself in a moral dilemma concerning YAML, I simply ask
   myself this: "What would Oren do?" ;)

Re: [Yaml-core] draft of YAML slideshow presentation

[Brian Dorsey <brian@do...>]

On Thu, Sep 19, 2002 at 12:23:40PM -0700, Brian Ingerson wrote:
> On 19/09/02 12:04 -0700, Brian Dorsey wrote:
> > 
> > -
> >     exif:
> >         ColorSpace: '1'
> >         ComponentsConfiguration: [ 1,2,3,0 ]
> 
> This probably won't parse the way you intend it. 
> 
>    ---
>    # This
>    - [ 1,2,3,0 ]
>    # Is the same as
>    -
>      - '1,2,3,0'
>    # not (what you intend)
>    -
>      - 1
>      - 2
>      - 3
>      - 0
> 
> 
> BTW, If you are parsing this with one of the public tools, you should let the
> author know.

Interesting... It seems to work great. ;) I'm using pyYAML, and the YAML
I posted was all generated by yaml.dump()

An interactive example of how this works for me:
Python 2.2.1 (#34, Apr  9 2002, 19:34:33) [MSC 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import yaml
>>> data = yaml.load(open(r"d:\photoarchive\metadata.txt",'r').read()).next()
>>> data[0]['exif']['ComponentsConfiguration']
[1, 2, 3, 0]
>>> type(data[0]['exif']['ComponentsConfiguration'])
<type 'list'>
>>>

So, it seems to be safely making the round-trip. 

I was happily suprised when the generated YAML came out using both "-"
notation and "[]" notation. I was under the impression that YAML allows
both the usual "-" and the collapsed "[]" to indicate sequences?

Take care,
-Brian

Re: [Yaml-core] draft of YAML slideshow presentation

[Brian Ingerson <ingy@tt...>]

On 19/09/02 23:29 +0300, Oren Ben-Kiki wrote:
> Neil Watkiss [mailto:neilw@...]  wrote:
> > Here's a bug:
> > 
> >     same:
> >       - !domain.tld,2002/type%30%10 value
> >       - !domain.tld,2002/type\0x30\n value
> >     different: # As far as the YAML parser is concerned
> >       - !domain.tld,2002/type0%10 value
> 
> 
> \-style escapes will be taken out of the next draft. That said, you are
> right - it should have been \x30.
> 
> > On a separate note, maybe we *should* support \0: I think it 
> > should mean a literal NULL byte, just like \z.
> 
> Perhaps we could just use \0 instead of \z. I don't know whether it is worth
> making a change, at this point anyway.

\z and friends were always a bit cutesy for me. What was the precedent for
these anyway?

Cheers, Brian

PS Oren, check inbox

Re: [Yaml-core] Dirty YAML

[Brian Ingerson <ingy@tt...>]

On 19/09/02 23:21 +0300, Oren Ben-Kiki wrote:
> Brian Ingerson [mailto:ingy@...] 
> > IMVHO most YAML applications will be implemented in a single 
> > language. That's just the way things happen. This is 
> > especially true with serialization. The unknown types 
> > proposal allows a given language to use "clean yaml" (no 
> > explicit types, no schema) and be able to round trip 100% 
> > within the usage of a specific loader/dumper. This comes at 
> > the expense of interoperability with other languages and/or 
> > loaders. Clean yaml drops to say 70-90%.
> > 
> > But never fear. That's where explicit typing and schemas come 
> > in. They provide the specifics "YAML types" that must be 
> > used. Of course, both loaders must not only know those types, 
> > but also obey the YAML semantics for those types for true 
> > interoperability. Of course this is called "dirty yaml". It's 
> > more work, and it's uglier, but you get stronger typing.
> 
> +1
> 
> > Sometimes you can get interoperability with clean yaml. You 
> > just need to tell your Perl dumper to format things in such a 
> > way that the Python Loader will get it right. As certain 
> > applications become common, I think we'll see more 
> > loader/dumpers offering compatibility modes.
> 
> I think this would be more common than you make it out to be. The "out of
> the box" interoperability would depend on how far the data is from a simple
> map/seq/str types. A great many cases, this won't be too far.
 
+1

Yeah. I totally agree. I guess I was playing too much the defender. There
will be many cases when "It Just Works".

My main point is that even though clean yaml isn't (nor can be) semantically
air-tight, we have plenty of other weapons, should you need them.

Cheers, Brian

RE: [Yaml-core] draft of YAML slideshow presentation

[Oren Ben-Kiki <orenbk@ri...>]

Neil Watkiss [mailto:neilw@...]  wrote:
> Here's a bug:
> 
>     same:
>       - !domain.tld,2002/type%30%10 value
>       - !domain.tld,2002/type\0x30\n value
>     different: # As far as the YAML parser is concerned
>       - !domain.tld,2002/type0%10 value


\-style escapes will be taken out of the next draft. That said, you are
right - it should have been \x30.

> On a separate note, maybe we *should* support \0: I think it 
> should mean a literal NULL byte, just like \z.

Perhaps we could just use \0 instead of \z. I don't know whether it is worth
making a change, at this point anyway.

Have fun,

	Oren Ben-Kiki

RE: [Yaml-core] Dirty YAML

[Oren Ben-Kiki <orenbk@ri...>]

Brian Ingerson [mailto:ingy@...] 
> IMVHO most YAML applications will be implemented in a single 
> language. That's just the way things happen. This is 
> especially true with serialization. The unknown types 
> proposal allows a given language to use "clean yaml" (no 
> explicit types, no schema) and be able to round trip 100% 
> within the usage of a specific loader/dumper. This comes at 
> the expense of interoperability with other languages and/or 
> loaders. Clean yaml drops to say 70-90%.
> 
> But never fear. That's where explicit typing and schemas come 
> in. They provide the specifics "YAML types" that must be 
> used. Of course, both loaders must not only know those types, 
> but also obey the YAML semantics for those types for true 
> interoperability. Of course this is called "dirty yaml". It's 
> more work, and it's uglier, but you get stronger typing.

+1

> Sometimes you can get interoperability with clean yaml. You 
> just need to tell your Perl dumper to format things in such a 
> way that the Python Loader will get it right. As certain 
> applications become common, I think we'll see more 
> loader/dumpers offering compatibility modes.

I think this would be more common than you make it out to be. The "out of
the box" interoperability would depend on how far the data is from a simple
map/seq/str types. A great many cases, this won't be too far.

Have fun,

	Oren Ben-Kiki

[Yaml-core] !include

[Brian Ingerson <ingy@tt...>]

I noticed that Steve has implemented a !!read transfer for PyYaml. I think
this is great. It's similar to the !include we proposed some time ago.
Steve's usage is:

    foo: !!read slides.yml

I was thinking a more generic approach would be:

    !include file-uri#ypath

This gives us flexibility in both what kinds of files we can use and what
parts of those files to use.

One problem IIRC, is that a file contains a stream rather than a document.
Should we make YPATH be able identify stream content? Also, what happens when
the ypath refers to more than one node?

Steve, what do you do if the !!read target contains a stream with more than
one document?

Cheers, Brian

Re: [Yaml-core] draft of YAML slideshow presentation

[Neil Watkiss <neilw@ActiveState.com>]

Oren Ben-Kiki [19/09/02 22:49 +0300]:
> This actually is not valid. We no longer allow a schema (it is always
> implicit). Also it is taguri, not tag. See the spec section for some
> examples.

Here's a bug:

    same:
      - !domain.tld,2002/type%30%10 value
      - !domain.tld,2002/type\0x30\n value
    different: # As far as the YAML parser is concerned
      - !domain.tld,2002/type0%10 value

/same/1 is not what you think it is. At least, I think it's not. libyaml
would get a parse error on it. If you want %30 you have to say \x30, not
\0x30. Otherwise "\0" is a parse error; we don't support the \0 escape.

On a separate note, maybe we *should* support \0: I think it should mean a
literal NULL byte, just like \z.

Either way, the example needs fixing.

Later,
Neil

[Yaml-core] Dirty YAML

[Brian Ingerson <ingy@tt...>]

It just struck me that Oren's "unknown types" proposal is very
complimentary to weakly typed languages like Perl, without dropping the
ball for more strongly typed languages.

Every programming language supports the same types of data (more or
less), but has a different take on the exact semantics of those types.
YAML is meant to serve many programming languages. By moving our
implicit type resolution to the loader level, we are allowing a
language's loader to serve itself in the best way possible.

IMVHO most YAML applications will be implemented in a single language. That's
just the way things happen. This is especially true with serialization. The
unknown types proposal allows a given language to use "clean yaml" (no
explicit types, no schema) and be able to round trip 100% within the usage of
a specific loader/dumper. This comes at the expense of interoperability with
other languages and/or loaders. Clean yaml drops to say 70-90%.

But never fear. That's where explicit typing and schemas come in. They
provide the specifics "YAML types" that must be used. Of course, both loaders
must not only know those types, but also obey the YAML semantics for those
types for true interoperability. Of course this is called "dirty yaml". It's
more work, and it's uglier, but you get stronger typing.

Sometimes you can get interoperability with clean yaml. You just need to tell
your Perl dumper to format things in such a way that the Python Loader will
get it right. As certain applications become common, I think we'll see more
loader/dumpers offering compatibility modes.

Another benefit of unknown types is that it keeps the libyaml parser clean.
Of course it moves the work downstream, and you might argue that this is
duplication. But that's my whole point. It's *not* duplication because each
language has differing types and semantics.

Cheers, Brian

RE: [Yaml-core] the seq-map syntax error

[Oren Ben-Kiki <orenbk@ri...>]

Brian Ingerson [mailto:ingy@...] wrote:
> > - But a YAML mapping embedded in a seq I'd prefer to write
> >   like:  this
> >   becase:  it explicitly shows the nesting
> 
> The above does not parse today, correct? Unquoted keys may 
> not span lines.

+1
 
> So if I'm reading the intent correctly, you'd want:
> 
>     - 'But a YAML mapping embedded in a seq I'd prefer to write
>       like':  this
>       becase:  it explicitly shows the nesting

Yes, that is valid today.

Mike is right, however:
> That is the normal way people write lists.
> 1. Point 1.
> 2. Point 2.

In YAML this would be:
- Point 1.
- Point 2.

Which is _not_ currently valid, but could be if we decide to allow it.

Have fun,

	Oren Ben-Kiki

RE: [Yaml-core] draft of YAML slideshow presentation

[Oren Ben-Kiki <orenbk@ri...>]

Brian Ingerson [mailto:ingy@...] wrote:
> Nice job. The only thing missing is any talk about type URIs. 
> You might want a simple exampe like this:
> 
>     --- # These are equivalent
>     - !tag:yaml.org,2002/int 42

This actually is not valid. We no longer allow a schema (it is always
implicit). Also it is taguri, not tag. See the spec section for some
examples.

Have fun,

	Oren Ben-Kiki

Re: [Yaml-core] draft of YAML slideshow presentation

[Brian Ingerson <ingy@tt...>]

On 19/09/02 12:04 -0700, Brian Dorsey wrote:
> On Thu, Sep 19, 2002 at 11:15:36AM -0400, Steve Howell wrote:
> > One more call for cool YAML examples--you West Coast YAMLers should be getting
> > out of bed now, send me your data!
> > 
> 
> Heh, you specified cool examples, and mine aren't. ;)
> 
> -
>     exif:
>         ColorSpace: '1'
>         ComponentsConfiguration: [ 1,2,3,0 ]

This probably won't parse the way you intend it. 

   ---
   # This
   - [ 1,2,3,0 ]
   # Is the same as
   -
     - '1,2,3,0'
   # not (what you intend)
   -
     - 1
     - 2
     - 3
     - 0

Cheers, Brian

BTW, If you are parsing this with one of the public tools, you should let the
author know.

Re: [Yaml-core] draft of YAML slideshow presentation

[Brian Ingerson <ingy@tt...>]

On 19/09/02 14:09 -0400, Steve Howell wrote:
> Brian's suggestions have now been incorporated in various places:

Nice job. The only thing missing is any talk about type URIs. You might want
a simple exampe like this:

    --- # These are equivalent
    - !tag:yaml.org,2002/int 42
    - !yaml.org,2002/int 42
    - !yaml.org/int 42
    - !int 42
    - 42       # usually

    --- # Language specific
    - !python.yaml.org/tuple [1, 2, 3]
    - !python/tuple [1, 2, 3]
    
    --- # Domain specific
    - !showell.org,2002/SlideShowell/^slide {}
    - !^bullet
    
> 
> http://mountainwebtools.com/SlideShowell/Slide-0.html
> 
> ----- Original Message ----- 
> From: "Brian Ingerson" <ingy@...>
> To: "Steve Howell" <showell@...>
> Cc: <yaml-core@...>
> Sent: Thursday, September 19, 2002 1:20 PM
> Subject: Re: [Yaml-core] draft of YAML slideshow presentation
> 
> 
> > On 19/09/02 11:15 -0400, Steve Howell wrote:
> > > Suggestions welcome:
> > 
> >   - Mention http://www.yaml.org
> >   - Show YAML inside YAML: |
> >      ---
> >      more: YAML
> >      no: escaping (like in XML)
> >   - You probably want to say something about typing. Implicit typing is
> >     pretty cool. At least show how integers are serialized in Python.
> >   ...
> > 
> > 
> > Here is some output from phone2yaml, a program that ships with YAML.pm, and
> > yamlizes the TSV export data from a palm pilot's address book.
> > 
> > ---
> > company: Baxter Cycle
> > phones:
> >   work: 712-781-2351
> > address:
> >   street: 400 Lincoln St
> >   city: Marne
> >   state: IA
> >   postal: 51552
> > ---
> > name:
> >   first: Tracy
> >   last: Bennett
> > title: Veternarian
> > company: Bird&Exotic Clinic of Seattle
> > phones:
> >   work: 206-783-4538
> >   home: 206-783-1679
> > address:
> >   street: 1546 NW Market st.
> >   city: Ballard
> >   state: Wa
> > 
> > You can mock up the data if you want. Note that the data is
> > pretty-printed to match the order of the palm pilot itself.
> > 
> > 
> > -------------------------------------------------------
> > This sf.net email is sponsored by:ThinkGeek
> > Welcome to geek heaven.
> > http://thinkgeek.com/sf
> > _______________________________________________
> > Yaml-core mailing list
> > Yaml-core@...
> > https://lists.sourceforge.net/lists/listinfo/yaml-core
> > 

Re: [Yaml-core] draft of YAML slideshow presentation

[Brian Dorsey <brian@do...>]

On Thu, Sep 19, 2002 at 11:15:36AM -0400, Steve Howell wrote:
> One more call for cool YAML examples--you West Coast YAMLers should be getting
> out of bed now, send me your data!
> 

Heh, you specified cool examples, and mine aren't. ;)

However, here is a bit of YAML I'm actually using to store metadata
about digital photos. I'm currently working on a simple system for
organizing my photos and this is my first datastore - I'm eventually
planning to support several options, but YAML seems like a good one to
start with. 

It's a ridiculously simple structure, but there is a lot of info per
photo, so it might not be all that useful as an example. The exif data
is embeded into the jpegs by the camera itself. 

Anyway, here is a snippet:
---
-
    exif:
        ColorSpace: '1'
        ComponentsConfiguration: [ 1,2,3,0 ]
        CompressedBitsPerPixel: '2/1'
        Compression: '6'
        DateTime: '2002:07:14 12:04:21'
        DateTimeDigitized: '2002:07:14 12:04:21'
        DateTimeOriginal: '2002:07:14 12:04:21'
        ExifImageHeight: '1200'
        ExifImageWidth: '1600'
        ExifVersion: [ 48,50,49,48 ]
        ExposureBiasValue: '0.0'
        ExposureProgram: Program Normal
        ExposureTime: '1/106'
        FNumber: '2.6'
        Flash: no
        FocalLength: '7.2'
        ISOSpeedRatings: '0'
        ImageDescription: ''
        Interoperability_IFD_Pointer: '886'
        JPEGInterchangeFormat: '2036'
        JPEGInterchangeFormatLength: '5042'
        LightSource: Unknown
        Make: NIKON
        MakerNote: [
78,105,107,111,110,0,1,0,11,0,2,0,2,0,6,0,0,0,38,4,0,0,3,0,3,0,1,0,0,0,11,0,0,0,4,0,3,0,1,0,0,0,2,0,0,0,5,0,3,0,1,0,0,0,0,0,0,0,6,0,3,0,1,0,0,0,0,0,0,0,7,0,3,0,1,0,0,0,0,0,0,0,8,0,5,0,1,0,0,0,44,4,0,0,9,0,2,0,20,0,0,0,52,4,0,0,10,0,5,0,1,0,0,0,72,4,0,0,11,0,3,0,1,0,0,0,0,0,0,0,0,15,4,0,30,0,0,0,80,4,0,0,0,0,0,0,48,56,46,48,48,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,100,0,0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0,0,0,36,250,19,160,0,0,0,0,36,149,0,0,20,65,0,0,36,149,0,0,20,65,0,0,0,0,0,0,0,0,10,206,0,0,0,0,0,0,16,132,17,90,16,232,15,126,4,198,11,145,0,0,0,0,20,24,10,0,0,0,0,0,2,0,0,0,0,4,64,0,67,32,0,0,74,3,20,90,16,1,126,0,11,95,100,0,2,228,16,13,16,97,124,0,255,0,89,109,50,12,176,248,0,0
]
        MaxApertureValue: '2.6'
        MeteringMode: Matrix
        Model: E950
        Orientation: '1'
        ResolutionUnit: Inch
        Software: 'v981-77'
        Tag0x9286: [
0,0,0,0,0,0,0,0,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32
]
        Tag0xa000: [ 48,49,48,48 ]
        Tag0xa300: [ 3 ]
        Tag0xa301: [ 1 ]
        XResolution: '300/1'
        YCbCrPositioning: '2'
        YResolution: '300/1'
    filebytes: 409898
    filemodifieddate: '2002/07/14 12:04:20'
    filename: '20020714-1204-21.JPG'
    filepath: 'd:\photoarchive\20020714\'
    filetype: JPG
    friendlyname: '20020714-1204-21'
-
    exif:
        ColorSpace: '1'
        ComponentsConfiguration: [ 1,2,3,0 ]
        CompressedBitsPerPixel: '2/1'
        Compression: '6'
        DateTime: '2002:07:14 12:04:33'
        DateTimeDigitized: '2002:07:14 12:04:33'
        DateTimeOriginal: '2002:07:14 12:04:33'
        ExifImageHeight: '1200'
        ExifImageWidth: '1600'
        ExifVersion: [ 48,50,49,48 ]
        ExposureBiasValue: '0.0'
        ExposureProgram: Program Normal
        ExposureTime: '1/45'
        FNumber: '4.4'
        Flash: no
        FocalLength: '7.2'
        ISOSpeedRatings: '0'
        ImageDescription: ''
        Interoperability_IFD_Pointer: '886'
        JPEGInterchangeFormat: '2036'
        JPEGInterchangeFormatLength: '4334'
        LightSource: Unknown
        Make: NIKON
        MakerNote: [
78,105,107,111,110,0,1,0,11,0,2,0,2,0,6,0,0,0,38,4,0,0,3,0,3,0,1,0,0,0,11,0,0,0,4,0,3,0,1,0,0,0,2,0,0,0,5,0,3,0,1,0,0,0,0,0,0,0,6,0,3,0,1,0,0,0,0,0,0,0,7,0,3,0,1,0,0,0,0,0,0,0,8,0,5,0,1,0,0,0,44,4,0,0,9,0,2,0,20,0,0,0,52,4,0,0,10,0,5,0,1,0,0,0,72,4,0,0,11,0,3,0,1,0,0,0,0,0,0,0,0,15,4,0,30,0,0,0,80,4,0,0,0,0,0,0,48,56,46,48,48,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,100,0,0,0,1,6,0,128,1,124,0,0,0,0,255,2,0,0,0,0,26,237,17,160,0,0,0,0,25,146,0,0,21,171,0,0,86,151,0,0,16,230,0,0,0,0,0,0,0,0,12,14,0,47,0,0,0,0,14,209,15,146,16,187,16,83,3,74,11,150,0,0,0,252,16,19,12,0,0,0,0,0,0,0,0,16,0,8,0,0,59,49,0,0,74,2,18,87,17,17,115,0,24,89,100,0,2,228,16,15,17,129,112,0,255,0,95,105,50,12,176,248,0,0
]
        MaxApertureValue: '2.6'
        MeteringMode: Matrix
        Model: E950
        Orientation: '1'
        ResolutionUnit: Inch
        Software: 'v981-77'
        Tag0x9286: [
0,0,0,0,0,0,0,0,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32
]
        Tag0xa000: [ 48,49,48,48 ]
        Tag0xa300: [ 3 ]
        Tag0xa301: [ 1 ]
        XResolution: '300/1'
        YCbCrPositioning: '2'
        YResolution: '300/1'
    filebytes: 355166
    filemodifieddate: '2002/07/14 12:04:32'
    filename: '20020714-1204-33.JPG'
    filepath: 'd:\photoarchive\20020714\'
    filetype: JPG
    friendlyname: '20020714-1204-33'
-
    exif:
        ColorSpace: '1'
        ComponentsConfiguration: [ 1,2,3,0 ]
        CompressedBitsPerPixel: '2/1'
        Compression: '6'
        DateTime: '2002:07:14 12:04:50'
        DateTimeDigitized: '2002:07:14 12:04:50'
        DateTimeOriginal: '2002:07:14 12:04:50'
        ExifImageHeight: '1200'
        ExifImageWidth: '1600'
        ExifVersion: [ 48,50,49,48 ]
        ExposureBiasValue: '0.0'
        ExposureProgram: Program Normal
        ExposureTime: '1/50'
        FNumber: '4.4'
        Flash: no
        FocalLength: '7.2'
        ISOSpeedRatings: '0'
        ImageDescription: ''
        Interoperability_IFD_Pointer: '886'
        JPEGInterchangeFormat: '2036'
        JPEGInterchangeFormatLength: '4660'
        LightSource: Unknown
        Make: NIKON
        MakerNote: [
78,105,107,111,110,0,1,0,11,0,2,0,2,0,6,0,0,0,38,4,0,0,3,0,3,0,1,0,0,0,11,0,0,0,4,0,3,0,1,0,0,0,2,0,0,0,5,0,3,0,1,0,0,0,0,0,0,0,6,0,3,0,1,0,0,0,0,0,0,0,7,0,3,0,1,0,0,0,0,0,0,0,8,0,5,0,1,0,0,0,44,4,0,0,9,0,2,0,20,0,0,0,52,4,0,0,10,0,5,0,1,0,0,0,72,4,0,0,11,0,3,0,1,0,0,0,0,0,0,0,0,15,4,0,30,0,0,0,80,4,0,0,0,0,0,0,48,56,46,48,48,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,100,0,0,0,1,24,0,128,1,112,0,0,0,0,255,2,0,0,0,0,30,114,17,106,0,0,0,0,28,55,0,0,22,59,0,0,77,54,0,0,20,243,0,0,0,0,0,0,0,0,12,14,0,47,0,0,0,0,5,79,6,46,6,237,6,123,23,255,11,202,0,0,0,0,20,17,12,0,0,0,0,0,0,0,4,0,0,0,0,0,47,34,0,0,70,2,22,72,13,49,78,0,21,90,100,0,2,228,14,15,13,241,79,0,255,0,76,96,50,12,176,248,0,0
]
        MaxApertureValue: '2.6'
        MeteringMode: Matrix
        Model: E950
        Orientation: '1'
        ResolutionUnit: Inch
        Software: 'v981-77'
        Tag0x9286: [
0,0,0,0,0,0,0,0,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32,32
]
        Tag0xa000: [ 48,49,48,48 ]
        Tag0xa300: [ 3 ]
        Tag0xa301: [ 1 ]
        XResolution: '300/1'
        YCbCrPositioning: '2'
        YResolution: '300/1'
    filebytes: 404699
    filemodifieddate: '2002/07/14 12:04:50'
    filename: '20020714-1204-50.JPG'
    filepath: 'd:\photoarchive\20020714\'
    filetype: JPG
    friendlyname: '20020714-1204-50'

Re: [Yaml-core] draft of YAML slideshow presentation

[Steve Howell <showell@zi...>]

Brian's suggestions have now been incorporated in various places:

http://mountainwebtools.com/SlideShowell/Slide-0.html

----- Original Message ----- 
From: "Brian Ingerson" <ingy@...>
To: "Steve Howell" <showell@...>
Cc: <yaml-core@...>
Sent: Thursday, September 19, 2002 1:20 PM
Subject: Re: [Yaml-core] draft of YAML slideshow presentation


> On 19/09/02 11:15 -0400, Steve Howell wrote:
> > Suggestions welcome:
> 
>   - Mention http://www.yaml.org
>   - Show YAML inside YAML: |
>      ---
>      more: YAML
>      no: escaping (like in XML)
>   - You probably want to say something about typing. Implicit typing is
>     pretty cool. At least show how integers are serialized in Python.
>   ...
> 
> 
> Here is some output from phone2yaml, a program that ships with YAML.pm, and
> yamlizes the TSV export data from a palm pilot's address book.
> 
> ---
> company: Baxter Cycle
> phones:
>   work: 712-781-2351
> address:
>   street: 400 Lincoln St
>   city: Marne
>   state: IA
>   postal: 51552
> ---
> name:
>   first: Tracy
>   last: Bennett
> title: Veternarian
> company: Bird&Exotic Clinic of Seattle
> phones:
>   work: 206-783-4538
>   home: 206-783-1679
> address:
>   street: 1546 NW Market st.
>   city: Ballard
>   state: Wa
> 
> You can mock up the data if you want. Note that the data is
> pretty-printed to match the order of the palm pilot itself.
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Yaml-core mailing list
> Yaml-core@...
> https://lists.sourceforge.net/lists/listinfo/yaml-core
> 

Re: [Yaml-core] the seq-map syntax error

[Brian Ingerson <ingy@tt...>]

On 19/09/02 09:59 -0700, Mike Orr wrote:
> On Thu, Sep 19, 2002 at 10:38:31AM +0300, Oren Ben-Kiki wrote:
> > Point 4: I found myself catching myself typing this (not further indenting
> > the '-') and then fixing the sequence entries. Clark was bitten by this as
> > well. It seems there is a strong human tendency is to consider '-' as part
> > of the indentation, regardless of _writing_ the spec. I can only speculate
> > that this tendency would be stronger in someone merely _reading_ it :-)
> 
> That is the normal way people write lists.
> ** Point 1.
> ** Point 2.
> 
> That is the normal way people write lists.
> 1. Point 1.
> 2. Point 2.
> 
> As opposed to this, which looks unnatural.
>   ** Point 1.
>   ** Point 2.
> 
> As opposed to this, which looks unnatural.
>   1. Point 1.
>   2. Point 2.
> 
> 
> 
> - But a YAML mapping embedded in a seq I'd prefer to write
>   like:  this
>   becase:  it explicitly shows the nesting

The above does not parse today, correct? Unquoted keys may not span lines.

So if I'm reading the intent correctly, you'd want:

    - 'But a YAML mapping embedded in a seq I'd prefer to write
      like':  this
      becase:  it explicitly shows the nesting

Cheers, Brian

[Yaml-core] what YAML Ain't

[Steve Howell <showell@zi...>]

http://wiki.yaml.org/yamlwiki/WhatYamlAint


Please add to this list, if you know of possible YAML alternatives.  A good way
to understand YAML is to understand what it isn't.

Re: [Yaml-core] draft of YAML slideshow presentation

[Brian Ingerson <ingy@tt...>]

On 19/09/02 11:15 -0400, Steve Howell wrote:
> Suggestions welcome:

  - Mention http://www.yaml.org
  - Show YAML inside YAML: |
     ---
     more: YAML
     no: escaping (like in XML)
  - You probably want to say something about typing. Implicit typing is
    pretty cool. At least show how integers are serialized in Python.
  ...


Here is some output from phone2yaml, a program that ships with YAML.pm, and
yamlizes the TSV export data from a palm pilot's address book.

---
company: Baxter Cycle
phones:
  work: 712-781-2351
address:
  street: 400 Lincoln St
  city: Marne
  state: IA
  postal: 51552
---
name:
  first: Tracy
  last: Bennett
title: Veternarian
company: Bird&Exotic Clinic of Seattle
phones:
  work: 206-783-4538
  home: 206-783-1679
address:
  street: 1546 NW Market st.
  city: Ballard
  state: Wa

You can mock up the data if you want. Note that the data is
pretty-printed to match the order of the palm pilot itself.

Re: [Yaml-core] the seq-map syntax error

[<iron@ms...>]

On Thu, Sep 19, 2002 at 10:38:31AM +0300, Oren Ben-Kiki wrote:
> Point 4: I found myself catching myself typing this (not further indenting
> the '-') and then fixing the sequence entries. Clark was bitten by this as
> well. It seems there is a strong human tendency is to consider '-' as part
> of the indentation, regardless of _writing_ the spec. I can only speculate
> that this tendency would be stronger in someone merely _reading_ it :-)

That is the normal way people write lists.
** Point 1.
** Point 2.

That is the normal way people write lists.
1. Point 1.
2. Point 2.

As opposed to this, which looks unnatural.
  ** Point 1.
  ** Point 2.

As opposed to this, which looks unnatural.
  1. Point 1.
  2. Point 2.



- But a YAML mapping embedded in a seq I'd prefer to write
  like:  this
  becase:  it explicitly shows the nesting

- Rather than
like:  this
because:  it looks like a list followed by a seq, whatever that means


Comments inside block scalars also seem hard to grok, and no other
programming language allows it, except perhaps those that use
"missing operator" to mean "concatenate" and also make you quote
each chunk of the scalar.  Do we have an example of where it would
be highly desirable over putting the comment after the scalar or
in another legal position?


Re spaces before comments, I and some other programmers have a
convention:
# This is a text comment, so it has a space preceding it.
#var = "This is a commented code line so there's no space before it."

But I have no objection to requiring the space in YAML, especially since
we allow unquoted strings and the other languages do not.  That in
itself argues for keeping the other rules more strict, to minimize
accidental misparsing.

-- 
-Mike (Iron) Orr, iron@...  (if mail problems: mso@...)
   http://iron.cx/     English * Esperanto * Russkiy * Deutsch * Espan~ol

Re: [Yaml-core] draft of YAML slideshow presentation

[Clark C. Evans <cce@cl...>]

I'll be on it tnoight.O

n Thu, Sep 19, 2002 at 11:15:36AM -0400, Steve Howell wrote:
| Suggestions welcome:
| 
| http://mountainwebtools.com/SlideShowell/Slide-0.html
| 
| Pardon the Python bias in these slides, but I am giving this talk to a bunch of
| Python programmers.
| 
| One more call for cool YAML examples--you West Coast YAMLers should be getting
| out of bed now, send me your data!
| 
| 
| 
| 
| 
| 
| -------------------------------------------------------
| This sf.net email is sponsored by:ThinkGeek
| Welcome to geek heaven.
| http://thinkgeek.com/sf
| _______________________________________________
| Yaml-core mailing list
| Yaml-core@...
| https://lists.sourceforge.net/lists/listinfo/yaml-core

[Yaml-core] draft of YAML slideshow presentation

[Steve Howell <showell@zi...>]

Suggestions welcome:

http://mountainwebtools.com/SlideShowell/Slide-0.html

Pardon the Python bias in these slides, but I am giving this talk to a bunch of
Python programmers.

One more call for cool YAML examples--you West Coast YAMLers should be getting
out of bed now, send me your data!

[Yaml-core] Why's RPC template proposal

[Steve Howell <showell@zi...>]

http://wiki.yaml.org/yamlwiki/WhysRpcTemplateProposal

All the content was in an earlier email; just wikifying it.

Cheers,

Steve

[Yaml-core] 2nd call for real world YAML samples

[Steve Howell <showell@zi...>]

If you have a good example of YAML being used in the real world, please send a
note to the mailing list.  Actual YAML examples would be the coolest, but it
would also be good to hear what you're working on.

Thanks,

Steve

RE: [Yaml-core] the seq-map syntax error

[Oren Ben-Kiki <orenbk@ri...>]

Steve Howell [mailto:showell@...] wrote:
> From: "Oren Ben-Kiki" <orenbk@...>
> > -   key: value
> >      value

*Blush*. I meant to write:

-   key: value
    valid: value

Somehow it got garbled.

> My intuition tells me that would be invalid.

No, it is valid.

> Please don't tell me it's something like this: [{'key': 
> 'value value'}]

Sorry, it is :-) Because the second "value" is more indented. It is the same
as:

This: value spans
 multiple lines

Have fun,

	Oren Ben-Kiki