Projects vs. Articles on CP
-
[ I originally posted this in the Soapbox where it went nowhere. Two folks suggested I post to the Lounge so here goes ... again. ] I'm seeing some articles that over a period of months and years have evolved into ongoing projects. First, let me try to distinguish an article from a project. The purpose of an article is to describe a specific technique, approach or solution to a problem. Except for corrections and minor revisions - it is fairly static in content and will eventually become obsolete over (a possibly long) time. Most CP articles fall into this generic category. A project is a complete working program or library. A CP article posted in this manner is more or less a "how-to" of how to use the program or library. As the code improves, the article text must be revised to correctly reflect new behavior. The article becomes a living document where any draft is a snapshot in time of where the project stood in its evolution. The big question: Is posting a project style article on CP a good thing? The trouble with projects hosted on CP is it is hard to see what was changed with each update. I seeing some of the same updated articles pop up over and over again. I can't tell if they are minor typo fixes, minor bug fixes, or contain significant new content (ie. is the article worth re-reading?). Might some of the project-articles fair better somewhere like SourceForge? One main benefit of SF is visibility of source revisions - including the article text - via the CVS or SVN repository feature of that site. I can easily see the value of a CP hosted "how to" article that links to an SF hosted project. Ideas?
-
[ I originally posted this in the Soapbox where it went nowhere. Two folks suggested I post to the Lounge so here goes ... again. ] I'm seeing some articles that over a period of months and years have evolved into ongoing projects. First, let me try to distinguish an article from a project. The purpose of an article is to describe a specific technique, approach or solution to a problem. Except for corrections and minor revisions - it is fairly static in content and will eventually become obsolete over (a possibly long) time. Most CP articles fall into this generic category. A project is a complete working program or library. A CP article posted in this manner is more or less a "how-to" of how to use the program or library. As the code improves, the article text must be revised to correctly reflect new behavior. The article becomes a living document where any draft is a snapshot in time of where the project stood in its evolution. The big question: Is posting a project style article on CP a good thing? The trouble with projects hosted on CP is it is hard to see what was changed with each update. I seeing some of the same updated articles pop up over and over again. I can't tell if they are minor typo fixes, minor bug fixes, or contain significant new content (ie. is the article worth re-reading?). Might some of the project-articles fair better somewhere like SourceForge? One main benefit of SF is visibility of source revisions - including the article text - via the CVS or SVN repository feature of that site. I can easily see the value of a CP hosted "how to" article that links to an SF hosted project. Ideas?
I agree on the significance of the difference.
rfmobile wrote:
Is posting a project style article on CP a good thing?
Yes, without a doubt in my opinion.
rfmobile wrote:
The trouble with projects hosted on CP is it is hard to see what was changed with each update.
rfmobile wrote:
Might some of the project-articles fair better somewhere like SourceForge?
No. I think what's needed is improved support within the CP framework for project type articles. I suspect once the current list of post upgrade issues have been sorted out Chris M and his team will likely be focussing on just this sort of thing. I suggest you come up with a spec of the kind of facilities projects would benefit from that are not needed by articles and either send it to Chris or if it's short then post it in the suggestions forum. You could even submit an article describing what you'd like with faked up screen shots :)
Nothing is exactly what it seems but everything with seems can be unpicked.
-
[ I originally posted this in the Soapbox where it went nowhere. Two folks suggested I post to the Lounge so here goes ... again. ] I'm seeing some articles that over a period of months and years have evolved into ongoing projects. First, let me try to distinguish an article from a project. The purpose of an article is to describe a specific technique, approach or solution to a problem. Except for corrections and minor revisions - it is fairly static in content and will eventually become obsolete over (a possibly long) time. Most CP articles fall into this generic category. A project is a complete working program or library. A CP article posted in this manner is more or less a "how-to" of how to use the program or library. As the code improves, the article text must be revised to correctly reflect new behavior. The article becomes a living document where any draft is a snapshot in time of where the project stood in its evolution. The big question: Is posting a project style article on CP a good thing? The trouble with projects hosted on CP is it is hard to see what was changed with each update. I seeing some of the same updated articles pop up over and over again. I can't tell if they are minor typo fixes, minor bug fixes, or contain significant new content (ie. is the article worth re-reading?). Might some of the project-articles fair better somewhere like SourceForge? One main benefit of SF is visibility of source revisions - including the article text - via the CVS or SVN repository feature of that site. I can easily see the value of a CP hosted "how to" article that links to an SF hosted project. Ideas?
rfmobile wrote:
I seeing some of the same updated articles pop up over and over again. I can't tell if they are minor typo fixes, minor bug fixes, or contain significant new content (ie. is the article worth re-reading?).
It's up to the author to describe the changes. With a more project oriented site like SourceForge, it's still up to the folks running the project, and there you'll also see that there's a huge lack in intelligent description of changes. If you want to check for code differences, download it and pop the new and old versions into your favorite diff app and check them out yourself. It's pretty much the same thing you'd have to do if you were using source control. In fact, there's nothing stopping you from source controlling the code to an article you yourself are interested in. I do that on occasion, if for no other reason than to keep a backup of the code, and it comes in handy if the author changes the code, because it lets me merge my changes with the author's changes.
rfmobile wrote:
Might some of the project-articles fair better somewhere like SourceForge?
Nope. SourceForge is not a community, it's a repository. There's a big difference. Marc
-
[ I originally posted this in the Soapbox where it went nowhere. Two folks suggested I post to the Lounge so here goes ... again. ] I'm seeing some articles that over a period of months and years have evolved into ongoing projects. First, let me try to distinguish an article from a project. The purpose of an article is to describe a specific technique, approach or solution to a problem. Except for corrections and minor revisions - it is fairly static in content and will eventually become obsolete over (a possibly long) time. Most CP articles fall into this generic category. A project is a complete working program or library. A CP article posted in this manner is more or less a "how-to" of how to use the program or library. As the code improves, the article text must be revised to correctly reflect new behavior. The article becomes a living document where any draft is a snapshot in time of where the project stood in its evolution. The big question: Is posting a project style article on CP a good thing? The trouble with projects hosted on CP is it is hard to see what was changed with each update. I seeing some of the same updated articles pop up over and over again. I can't tell if they are minor typo fixes, minor bug fixes, or contain significant new content (ie. is the article worth re-reading?). Might some of the project-articles fair better somewhere like SourceForge? One main benefit of SF is visibility of source revisions - including the article text - via the CVS or SVN repository feature of that site. I can easily see the value of a CP hosted "how to" article that links to an SF hosted project. Ideas?
A whole bunch of updates is not good in either case. I prefer articles that contain a working solution that encapsulates the ideas presented in the article. Rather than just "here's what you need to do to solve problem X", I prefer "here's the widget I created to solve problem X, how I created it, and how to use it". The scope of the article and widget should be narrow enough that only bug-fixes and corrections are required.
-
rfmobile wrote:
I seeing some of the same updated articles pop up over and over again. I can't tell if they are minor typo fixes, minor bug fixes, or contain significant new content (ie. is the article worth re-reading?).
It's up to the author to describe the changes. With a more project oriented site like SourceForge, it's still up to the folks running the project, and there you'll also see that there's a huge lack in intelligent description of changes. If you want to check for code differences, download it and pop the new and old versions into your favorite diff app and check them out yourself. It's pretty much the same thing you'd have to do if you were using source control. In fact, there's nothing stopping you from source controlling the code to an article you yourself are interested in. I do that on occasion, if for no other reason than to keep a backup of the code, and it comes in handy if the author changes the code, because it lets me merge my changes with the author's changes.
rfmobile wrote:
Might some of the project-articles fair better somewhere like SourceForge?
Nope. SourceForge is not a community, it's a repository. There's a big difference. Marc
Marc Clifton wrote:
If you want to check for code differences, download it and pop the new and old versions into your favorite diff app and check them out yourself. It's pretty much the same thing you'd have to do if you were using source control.
AFAIK, CP does not provide prior revisions. That's the problem. Perhaps a wiki-like revision mechanism is the answer. You can only see the current revision. I'd have to make a copy of the article every time it is updated - in anticipation of *possibly* needing to see the deltas. Most articles I read because they are interesting ... not because I have a specific need. My needs change. Weeks or months later I see an update on a past article whose content may be relevant to a new work assignment. I'd like to know whether the update is new significant content or corrections to a handful of typos.
-
Marc Clifton wrote:
If you want to check for code differences, download it and pop the new and old versions into your favorite diff app and check them out yourself. It's pretty much the same thing you'd have to do if you were using source control.
AFAIK, CP does not provide prior revisions. That's the problem. Perhaps a wiki-like revision mechanism is the answer. You can only see the current revision. I'd have to make a copy of the article every time it is updated - in anticipation of *possibly* needing to see the deltas. Most articles I read because they are interesting ... not because I have a specific need. My needs change. Weeks or months later I see an update on a past article whose content may be relevant to a new work assignment. I'd like to know whether the update is new significant content or corrections to a handful of typos.
rfmobile wrote:
AFAIK, CP does not provide prior revisions. That's the problem. Perhaps a wiki-like revision mechanism is the answer.
As I already said, put the code into source control. Your source control. And as I already said, if you have an existing copy of the code, even if it's not in source control, you can download the new version and do a diff on the files.
rfmobile wrote:
I'd have to make a copy of the article every time it is updated - in anticipation of *possibly* needing to see the deltas.
Well, duh. Or put it into some source control that you maintain. Honestly, this isn't that difficult to do yourself. In fact, you could probably write an app to automate a lot of the process. Marc
-
rfmobile wrote:
AFAIK, CP does not provide prior revisions. That's the problem. Perhaps a wiki-like revision mechanism is the answer.
As I already said, put the code into source control. Your source control. And as I already said, if you have an existing copy of the code, even if it's not in source control, you can download the new version and do a diff on the files.
rfmobile wrote:
I'd have to make a copy of the article every time it is updated - in anticipation of *possibly* needing to see the deltas.
Well, duh. Or put it into some source control that you maintain. Honestly, this isn't that difficult to do yourself. In fact, you could probably write an app to automate a lot of the process. Marc
Marc Clifton wrote:
Honestly, this isn't that difficult to do yourself.
Marc; You're right. This is easy to do manually for one or two articles.
Marc Clifton wrote:
In fact, you could probably write an app to automate a lot of the process.
If I have to write an application to supplement the website's own interface, what does that say about the usability of this site? Writing a "screen scraper" application is perilous work. One small theme change can break your code. Talk about brittle. Automated or not - would this be practical for a hundred or more articles? Is it desirable for 100 or more folks like my self to incur this extra work (maintaining our own private repository of select past articles) when the website *might* more easily be able to provide this as a feature? Judging from the mild response - this isn't a big deal for most folks. I humbly suggest you missed the point. I'll take the blame for that. Allow me to set it up this way ... 1. Article appears in the top 10 new/updated list today. 2. Article was updated so at least one prior revision exists - somewhere. 3. Article was originally submitted in 2002. The article isn't new - but is new to me. I either didn't see or have an interest in the article at the time it was written. Now I do have an interest in the article - including past revisions of both article text and source files. Your suggestion requires archiving all articles of potential interest a priori. That doesn't work for the above example. I appreciate your input. I infer from the small number of replies that this isn't a "must have" feature for this site. Someone might write a Firefox plug-in to do exactly that - or not. I've enjoyed this site all these years without such a feature so I'm sure I'll survive.
-
Marc Clifton wrote:
Honestly, this isn't that difficult to do yourself.
Marc; You're right. This is easy to do manually for one or two articles.
Marc Clifton wrote:
In fact, you could probably write an app to automate a lot of the process.
If I have to write an application to supplement the website's own interface, what does that say about the usability of this site? Writing a "screen scraper" application is perilous work. One small theme change can break your code. Talk about brittle. Automated or not - would this be practical for a hundred or more articles? Is it desirable for 100 or more folks like my self to incur this extra work (maintaining our own private repository of select past articles) when the website *might* more easily be able to provide this as a feature? Judging from the mild response - this isn't a big deal for most folks. I humbly suggest you missed the point. I'll take the blame for that. Allow me to set it up this way ... 1. Article appears in the top 10 new/updated list today. 2. Article was updated so at least one prior revision exists - somewhere. 3. Article was originally submitted in 2002. The article isn't new - but is new to me. I either didn't see or have an interest in the article at the time it was written. Now I do have an interest in the article - including past revisions of both article text and source files. Your suggestion requires archiving all articles of potential interest a priori. That doesn't work for the above example. I appreciate your input. I infer from the small number of replies that this isn't a "must have" feature for this site. Someone might write a Firefox plug-in to do exactly that - or not. I've enjoyed this site all these years without such a feature so I'm sure I'll survive.
rfmobile wrote:
If I have to write an application to supplement the website's own interface, what does that say about the usability of this site?
It says that the intent, goals, design, and decision making process for this site doesn't support that feature. I'd like Word to read my mind, what does that say about Word that I have to write my own mind reading software?
rfmobile wrote:
Writing a "screen scraper" application is perilous work.
I'm not talking screen scraper. I'm talking about something that automatically unzips a download into a folder, checks it into a repository with the same name as zip file, then does a diff and reports to you in nicely formatted HTML the changes the programmer made.
rfmobile wrote:
Now I do have an interest in the article - including past revisions of both article text and source files.
Ah, that's true. You wouldn't get the past revisions until you started the process. Out of curiosity, why the interest in past revisions? Do you look at change history in the repositories on Source Forge? If so, why? Marc
-
rfmobile wrote:
If I have to write an application to supplement the website's own interface, what does that say about the usability of this site?
It says that the intent, goals, design, and decision making process for this site doesn't support that feature. I'd like Word to read my mind, what does that say about Word that I have to write my own mind reading software?
rfmobile wrote:
Writing a "screen scraper" application is perilous work.
I'm not talking screen scraper. I'm talking about something that automatically unzips a download into a folder, checks it into a repository with the same name as zip file, then does a diff and reports to you in nicely formatted HTML the changes the programmer made.
rfmobile wrote:
Now I do have an interest in the article - including past revisions of both article text and source files.
Ah, that's true. You wouldn't get the past revisions until you started the process. Out of curiosity, why the interest in past revisions? Do you look at change history in the repositories on Source Forge? If so, why? Marc
Marc Clifton wrote:
I'd like Word to read my mind, what does that say about Word that I have to write my own mind reading software?
In the case of Word (and any other Microsoft product), I'd consider that a feature. M$ is already far too ingrained into my business as it stands. There's a joke in here somewhere about the new unannounced Microsoft Mind Reading Server. Okay ... enough kidding.
Marc Clifton wrote:
'm not talking screen scraper.
I'm interested in seeing "deltas" for both code and article text so that means scraping the article text from the HTML. I'm guessing you'd have to also scrape the source code download URL from the article text too.
Marc Clifton wrote:
Out of curiosity, why the interest in past revisions?
Here's an extreme example: someone publishes an article and then revokes the entire article. I might get lucky and find an older draft in the Google cache. A more common example: I download the code for an article, play with it a bit and then delete it because I don't have an immediate need. Months or years later - I see the article updated. What changed? Sometimes I can get a clue from the article comments but nothing exact. For sf.net, access to past revisions is more obvious in the context of regression. Latest version "x" exhibits some bug but older revision "y" doesn't.
modified on Monday, January 14, 2008 10:06:42 PM