www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Tangobos positioning

reply Bill Baxter <dnewsgroup billbaxter.com> writes:
It's great that the Tango team is bundling Tangobos with Tango these 
days.  But the Tangobos page on dsource[1] still basically says it's a 
band-aid only intended to help ease the transition from Phobos to Tango. 
  It takes the tone that Phobos is crap and you're going to "change 
sides" once you get your superior library working.

That's very different from the story I hear Kris preaching around here 
lately.  Which is: with Tangobos you can have the best of both worlds.

Also, just a marketing thing, but I would suggest that the wording on 
the Tangobos site be changed from emphasizing "compatibility layer" to 
"a port of Phobos".  It mentions in passing that it is basically a port, 
but I would make that the primary description.  The reason I think the 
"compatibility layer" description is bad is because "layer" implies slow 
and bug-prone.   Slow because it implies theres some extra layer of 
indirection there translating Phobos calls into Tango calls (which there 
isn't for 99% of it IIUC), and bug prone because trying to emulate API Y 
using API X usually reveals a number of "impedance mismatches" in 
practice.  It's a port of Phobos.  Or you could say a copy of Phobos 
with minor adjustments to make the code work with Tango.  "compatibility 
layer" sounds bloated.

Also the description paints the relationship between Tangobos and Tango 
as an uneasy one, which I don't think is the case (at least not anymore).


--bb
[1] http://www.dsource.org/projects/tangobos/
Jan 27 2008
parent reply "Kris" <foo bar.com> writes:
"Bill Baxter" <dnewsgroup billbaxter.com> wrote in message 
news:fniheh$1a0$1 digitalmars.com...
 It's great that the Tango team is bundling Tangobos with Tango these days. 
 But the Tangobos page on dsource[1] still basically says it's a band-aid 
 only intended to help ease the transition from Phobos to Tango. It takes 
 the tone that Phobos is crap and you're going to "change sides" once you 
 get your superior library working.

Then perhaps you could help to resolve that? There's always room for another willing pair of hands, and the doc is in a Wiki so that anyone and everyone can easily participate ;)
 Also, just a marketing thing, but I would suggest that the wording on the 
 Tangobos site be changed from emphasizing "compatibility layer" to "a port 
 of Phobos".  It mentions in passing that it is basically a port, but I 
 would make that the primary description.  The reason I think the 
 "compatibility layer" description is bad is because "layer" implies slow 
 and bug-prone.   Slow because it implies theres some extra layer of 
 indirection there translating Phobos calls into Tango calls (which there 
 isn't for 99% of it IIUC), and bug prone because trying to emulate API Y 
 using API X usually reveals a number of "impedance mismatches" in 
 practice.  It's a port of Phobos.  Or you could say a copy of Phobos with 
 minor adjustments to make the code work with Tango.  "compatibility layer" 
 sounds bloated.

Seems like a good approach. Would you mind fixing the Wiki page, please? That would be a big help
Jan 27 2008
parent reply Bill Baxter <dnewsgroup billbaxter.com> writes:
Kris wrote:
 "Bill Baxter" <dnewsgroup billbaxter.com> wrote in message 
 news:fniheh$1a0$1 digitalmars.com...
 It's great that the Tango team is bundling Tangobos with Tango these days. 
 But the Tangobos page on dsource[1] still basically says it's a band-aid 
 only intended to help ease the transition from Phobos to Tango. It takes 
 the tone that Phobos is crap and you're going to "change sides" once you 
 get your superior library working.

Then perhaps you could help to resolve that? There's always room for another willing pair of hands, and the doc is in a Wiki so that anyone and everyone can easily participate ;)
 Also, just a marketing thing, but I would suggest that the wording on the 
 Tangobos site be changed from emphasizing "compatibility layer" to "a port 
 of Phobos".  It mentions in passing that it is basically a port, but I 
 would make that the primary description.  The reason I think the 
 "compatibility layer" description is bad is because "layer" implies slow 
 and bug-prone.   Slow because it implies theres some extra layer of 
 indirection there translating Phobos calls into Tango calls (which there 
 isn't for 99% of it IIUC), and bug prone because trying to emulate API Y 
 using API X usually reveals a number of "impedance mismatches" in 
 practice.  It's a port of Phobos.  Or you could say a copy of Phobos with 
 minor adjustments to make the code work with Tango.  "compatibility layer" 
 sounds bloated.

Seems like a good approach. Would you mind fixing the Wiki page, please? That would be a big help

Sure, if it is ok for me to do so. It's still basically Gregor's page though, isn't it? You sure it's ok to rewrite it? --bb
Jan 27 2008
next sibling parent reply "Kris" <foo bar.com> writes:
"Bill Baxter" <dnewsgroup billbaxter.com> wrote in message 
news:fnimhu$cs6$1 digitalmars.com...
 Kris wrote:
 "Bill Baxter" <dnewsgroup billbaxter.com> wrote in message 
 news:fniheh$1a0$1 digitalmars.com...
 It's great that the Tango team is bundling Tangobos with Tango these 
 days. But the Tangobos page on dsource[1] still basically says it's a 
 band-aid only intended to help ease the transition from Phobos to Tango. 
 It takes the tone that Phobos is crap and you're going to "change sides" 
 once you get your superior library working.

Then perhaps you could help to resolve that? There's always room for another willing pair of hands, and the doc is in a Wiki so that anyone and everyone can easily participate ;)
 Also, just a marketing thing, but I would suggest that the wording on 
 the Tangobos site be changed from emphasizing "compatibility layer" to 
 "a port of Phobos".  It mentions in passing that it is basically a port, 
 but I would make that the primary description.  The reason I think the 
 "compatibility layer" description is bad is because "layer" implies slow 
 and bug-prone.   Slow because it implies theres some extra layer of 
 indirection there translating Phobos calls into Tango calls (which there 
 isn't for 99% of it IIUC), and bug prone because trying to emulate API Y 
 using API X usually reveals a number of "impedance mismatches" in 
 practice.  It's a port of Phobos.  Or you could say a copy of Phobos 
 with minor adjustments to make the code work with Tango.  "compatibility 
 layer" sounds bloated.

Seems like a good approach. Would you mind fixing the Wiki page, please? That would be a big help

Sure, if it is ok for me to do so. It's still basically Gregor's page though, isn't it? You sure it's ok to rewrite it?

hehe ... yeah, he may or may not appreciate that :) Here's a dedicated page in the Tango Wiki instead: http://dsource.org/projects/tango/wiki/TangobosInfo I understand you need to have a dsource user-id to access the wiki. If you can hook up on the irc channels, it might be simpler to set things up (we need to white-list your id, for example).
Jan 27 2008
parent Bill Baxter <dnewsgroup billbaxter.com> writes:
Kris wrote:
 "Bill Baxter" <dnewsgroup billbaxter.com> wrote in message 
 news:fnimhu$cs6$1 digitalmars.com...
 Kris wrote:


 Sure, if it is ok for me to do so.  It's still basically Gregor's page 
 though, isn't it?  You sure it's ok to rewrite it?

hehe ... yeah, he may or may not appreciate that :)

Ok then. Won't do that. I thought it was possible you and Gregor might have made some agreement that the Tango team was basically "in charge" of Tangobos now.
 Here's a dedicated page in the Tango Wiki instead: 
 http://dsource.org/projects/tango/wiki/TangobosInfo
 
 I understand you need to have a dsource user-id to access the wiki. If you 
 can hook up on the irc channels, it might be simpler to set things up (we 
 need to white-list your id, for example). 

No probs. Looks like I'm already on the "white-list". --bb
Jan 27 2008
prev sibling parent reply Gregor Richards <Richards codu.org> writes:
Bill Baxter wrote:
 Kris wrote:
 "Bill Baxter" <dnewsgroup billbaxter.com> wrote in message 
 news:fniheh$1a0$1 digitalmars.com...
 It's great that the Tango team is bundling Tangobos with Tango these 
 days. But the Tangobos page on dsource[1] still basically says it's a 
 band-aid only intended to help ease the transition from Phobos to 
 Tango. It takes the tone that Phobos is crap and you're going to 
 "change sides" once you get your superior library working.

Then perhaps you could help to resolve that? There's always room for another willing pair of hands, and the doc is in a Wiki so that anyone and everyone can easily participate ;)
 Also, just a marketing thing, but I would suggest that the wording on 
 the Tangobos site be changed from emphasizing "compatibility layer" 
 to "a port of Phobos".  It mentions in passing that it is basically a 
 port, but I would make that the primary description.  The reason I 
 think the "compatibility layer" description is bad is because "layer" 
 implies slow and bug-prone.   Slow because it implies theres some 
 extra layer of indirection there translating Phobos calls into Tango 
 calls (which there isn't for 99% of it IIUC), and bug prone because 
 trying to emulate API Y using API X usually reveals a number of 
 "impedance mismatches" in practice.  It's a port of Phobos.  Or you 
 could say a copy of Phobos with minor adjustments to make the code 
 work with Tango.  "compatibility layer" sounds bloated.

Seems like a good approach. Would you mind fixing the Wiki page, please? That would be a big help

Sure, if it is ok for me to do so. It's still basically Gregor's page though, isn't it? You sure it's ok to rewrite it? --bb

I would be downright /thrilled/ if somebody took that page from me. - Gregor Richards
Jan 27 2008
parent reply Bill Baxter <dnewsgroup billbaxter.com> writes:
Gregor Richards wrote:
 Bill Baxter wrote:
 Kris wrote:
 "Bill Baxter" <dnewsgroup billbaxter.com> wrote in message 
 news:fniheh$1a0$1 digitalmars.com...
 It's great that the Tango team is bundling Tangobos with Tango these 
 days. But the Tangobos page on dsource[1] still basically says it's 
 a band-aid only intended to help ease the transition from Phobos to 
 Tango. It takes the tone that Phobos is crap and you're going to 
 "change sides" once you get your superior library working.

Then perhaps you could help to resolve that? There's always room for another willing pair of hands, and the doc is in a Wiki so that anyone and everyone can easily participate ;)
 Also, just a marketing thing, but I would suggest that the wording 
 on the Tangobos site be changed from emphasizing "compatibility 
 layer" to "a port of Phobos".  It mentions in passing that it is 
 basically a port, but I would make that the primary description.  
 The reason I think the "compatibility layer" description is bad is 
 because "layer" implies slow and bug-prone.   Slow because it 
 implies theres some extra layer of indirection there translating 
 Phobos calls into Tango calls (which there isn't for 99% of it 
 IIUC), and bug prone because trying to emulate API Y using API X 
 usually reveals a number of "impedance mismatches" in practice.  
 It's a port of Phobos.  Or you could say a copy of Phobos with minor 
 adjustments to make the code work with Tango.  "compatibility layer" 
 sounds bloated.

Seems like a good approach. Would you mind fixing the Wiki page, please? That would be a big help

Sure, if it is ok for me to do so. It's still basically Gregor's page though, isn't it? You sure it's ok to rewrite it? --bb

I would be downright /thrilled/ if somebody took that page from me. - Gregor Richards

Ok. Right-o then. In the mean time I wrote this here: http://dsource.org/projects/tango/wiki/TangobosInfo I have no idea if any of it is true, but at least it is what I *wish* were true. Please check over it for factual errors if you have a sec. Thanks. --bb
Jan 27 2008
parent reply Lars Ivar Igesund <larsivar igesund.net> writes:
Bill Baxter wrote:

 Gregor Richards wrote:
 Bill Baxter wrote:
 Kris wrote:
 "Bill Baxter" <dnewsgroup billbaxter.com> wrote in message
 news:fniheh$1a0$1 digitalmars.com...
 It's great that the Tango team is bundling Tangobos with Tango these
 days. But the Tangobos page on dsource[1] still basically says it's
 a band-aid only intended to help ease the transition from Phobos to
 Tango. It takes the tone that Phobos is crap and you're going to
 "change sides" once you get your superior library working.

Then perhaps you could help to resolve that? There's always room for another willing pair of hands, and the doc is in a Wiki so that anyone and everyone can easily participate ;)
 Also, just a marketing thing, but I would suggest that the wording
 on the Tangobos site be changed from emphasizing "compatibility
 layer" to "a port of Phobos".  It mentions in passing that it is
 basically a port, but I would make that the primary description.
 The reason I think the "compatibility layer" description is bad is
 because "layer" implies slow and bug-prone.   Slow because it
 implies theres some extra layer of indirection there translating
 Phobos calls into Tango calls (which there isn't for 99% of it
 IIUC), and bug prone because trying to emulate API Y using API X
 usually reveals a number of "impedance mismatches" in practice.
 It's a port of Phobos.  Or you could say a copy of Phobos with minor
 adjustments to make the code work with Tango.  "compatibility layer"
 sounds bloated.

Seems like a good approach. Would you mind fixing the Wiki page, please? That would be a big help

Sure, if it is ok for me to do so. It's still basically Gregor's page though, isn't it? You sure it's ok to rewrite it? --bb

I would be downright /thrilled/ if somebody took that page from me. - Gregor Richards

Ok. Right-o then.

You're also added as a dev in case you would like to try your hand at remaining incompatibilities. -- Lars Ivar Igesund blog at http://larsivi.net DSource, #d.tango & #D: larsivi Dancing the Tango
Jan 27 2008
parent reply Bill Baxter <dnewsgroup billbaxter.com> writes:
Lars Ivar Igesund wrote:
 Bill Baxter wrote:
 
 Gregor Richards wrote:
 Bill Baxter wrote:
 Kris wrote:
 "Bill Baxter" <dnewsgroup billbaxter.com> wrote in message
 news:fniheh$1a0$1 digitalmars.com...
 It's great that the Tango team is bundling Tangobos with Tango these
 days. But the Tangobos page on dsource[1] still basically says it's
 a band-aid only intended to help ease the transition from Phobos to
 Tango. It takes the tone that Phobos is crap and you're going to
 "change sides" once you get your superior library working.

another willing pair of hands, and the doc is in a Wiki so that anyone and everyone can easily participate ;)
 Also, just a marketing thing, but I would suggest that the wording
 on the Tangobos site be changed from emphasizing "compatibility
 layer" to "a port of Phobos".  It mentions in passing that it is
 basically a port, but I would make that the primary description.
 The reason I think the "compatibility layer" description is bad is
 because "layer" implies slow and bug-prone.   Slow because it
 implies theres some extra layer of indirection there translating
 Phobos calls into Tango calls (which there isn't for 99% of it
 IIUC), and bug prone because trying to emulate API Y using API X
 usually reveals a number of "impedance mismatches" in practice.
 It's a port of Phobos.  Or you could say a copy of Phobos with minor
 adjustments to make the code work with Tango.  "compatibility layer"
 sounds bloated.

please? That would be a big help

though, isn't it? You sure it's ok to rewrite it? --bb

- Gregor Richards


You're also added as a dev in case you would like to try your hand at remaining incompatibilities.

Is there a list of those anywhere? (remaining incompatibilities, I mean) --bb
Jan 27 2008
parent reply Lars Ivar Igesund <larsivar igesund.net> writes:
Bill Baxter wrote:

 Lars Ivar Igesund wrote:
 Bill Baxter wrote:
 
 Gregor Richards wrote:
 Bill Baxter wrote:
 Kris wrote:
 "Bill Baxter" <dnewsgroup billbaxter.com> wrote in message
 news:fniheh$1a0$1 digitalmars.com...
 It's great that the Tango team is bundling Tangobos with Tango these
 days. But the Tangobos page on dsource[1] still basically says it's
 a band-aid only intended to help ease the transition from Phobos to
 Tango. It takes the tone that Phobos is crap and you're going to
 "change sides" once you get your superior library working.

another willing pair of hands, and the doc is in a Wiki so that anyone and everyone can easily participate ;)
 Also, just a marketing thing, but I would suggest that the wording
 on the Tangobos site be changed from emphasizing "compatibility
 layer" to "a port of Phobos".  It mentions in passing that it is
 basically a port, but I would make that the primary description.
 The reason I think the "compatibility layer" description is bad is
 because "layer" implies slow and bug-prone.   Slow because it
 implies theres some extra layer of indirection there translating
 Phobos calls into Tango calls (which there isn't for 99% of it
 IIUC), and bug prone because trying to emulate API Y using API X
 usually reveals a number of "impedance mismatches" in practice.
 It's a port of Phobos.  Or you could say a copy of Phobos with minor
 adjustments to make the code work with Tango.  "compatibility layer"
 sounds bloated.

please? That would be a big help

though, isn't it? You sure it's ok to rewrite it? --bb

- Gregor Richards


You're also added as a dev in case you would like to try your hand at remaining incompatibilities.

Is there a list of those anywhere? (remaining incompatibilities, I mean)

Beyond std.signals and std.thread (which will be hard to fix), I don't know of any. Actual users may be able to provide better answers. -- Lars Ivar Igesund blog at http://larsivi.net DSource, #d.tango & #D: larsivi Dancing the Tango
Jan 28 2008
parent reply Christopher Wright <dhasenan gmail.com> writes:
Lars Ivar Igesund wrote:
 Beyond std.signals and std.thread (which will be hard to fix), I don't know
 of any. Actual users may be able to provide better answers.
 

std.signals will be easy to fix, actually, since Tango has those gc notify methods. A week ago it was impossible to fix.
Jan 28 2008
parent Lars Ivar Igesund <larsivar igesund.net> writes:
Christopher Wright wrote:

 Lars Ivar Igesund wrote:
 Beyond std.signals and std.thread (which will be hard to fix), I don't
 know of any. Actual users may be able to provide better answers.
 

std.signals will be easy to fix, actually, since Tango has those gc notify methods. A week ago it was impossible to fix.

I know, the hard to fix comment was meant for std.thread only. -- Lars Ivar Igesund blog at http://larsivi.net DSource, #d.tango & #D: larsivi Dancing the Tango
Jan 28 2008