Yes, but the value in sites like/. is *EDITING*. They should filter shit like this so their s/n ratio is high, and thus their feed is valuable. This will bring subscribers and thus attention and thus ad impressions. At present, slashdot posts 90% crap found on other sites 4-9 days ago, and 7% crap like this and 3% interesting stuff, mostly hard-science stuff I'd find if I was still in school or in a non-CS/Real-Science job.
Go do volunteer basic computer literacy session for your local senior center. Don't try to convert them to linux or get them using Firefox or anything dumb like that. Just ask what their problems are, and how you can help. You will quickly understand how broken and unintuitive computer software is.
JAVA will be removed from the *NASDAQ-100 composite index*, but will continue to trade as normal until the company is actually acquired. This point was even mentioned in the press release, so extra points for getting it so (so!) basically wrong.
(Man,/. just continues to accumulate fail. I wonder when it'll implode.)
Funny, it continues to work really well for me every week as I get incremental ugprades that keep my system functional, secure and quite current. I'm more regularly up to date than my big-bang-distro friends, certainly with respect to security issues. I don't have to devote weekends to "upgrading the server", crossing my fingers to make sure it works. And I have multiple options with respect to handling libraries and development packages, and a clear path to mixing stable and unstable packages, and tweaking them as need be. Oh, if I want to add (or remove) support for a feature (e.g., bluetooth), I can easily rebuild every package that provides support for that feature... without getting the cruft of it when I don't need it.
Quite a "failure", indeed.
(To be clear: Gentoo does have social and architectural problems. But it does have some very compelling points, and the distro as a whole does keep moving on, quite usefully.)
a/ wanting a lot of new, varied information b/ not caring about skimming through a bunch of noise to pick out some signal.
A good aggregator makes the second point simpler, but the *varied* part of the former is interesting. It's not about "news" necessarily. I gave the Netflix new releases example, and I just now added the Project Gutenberg newly-available ebooks feed. Some from the forums for a couple of hobbies. A couple of my feeds are new software updates. Multiple feeds of the recent changes of wiki articles: either wikipedia pages I want to make sure maintain correctness, or the GnuCash wiki where I want to see if *anything* changes. Some are (semi-)personal blogs of friends where I just want to see what changes in their lives. Gentoo Linux News, CERT Advisories, new RFCs,...
We can easily see the day when common devices -- refridgerators, washing machines, tivo, &c. will have a feed for status. That gets into the filtering and intelligent aggregation aspect, and into The Future, but it's the direction we're headed.
328 is a lot of feeds, granted. But I think the tipping point of "just open them in a bunch of tabs" vs. "aggregator" is around 30 sites. If only because at that point, the tab widths are too small to see which site is which.:) Also, the time wasted in checking the sites that didn't have any changes becomes too great (depending, of course, on how often Your Favorite Sites change).
What I don't get about RSS is that it doesn't seem to be any better than, say, just setting up the sites I want to get news from on a bookmark folder in Firefox and middle clicking to open them all in new tabs.
I did that too, about 4 years ago. Then it grew to opening up 30 tabs, and having to a) recall what the previous web-page state was and b) identify if anything changed in a mix of differing, sometimes slow-to-load layouts. It was in competition with my slashboxes, which were quite novel and useful at the time. It was great, but then I discovered RSS.
My current (bloglines) subscription list is 327 feeds. They range far and wide: tech commentary, comics, raw news, politics, beer/homebrewing, trashy entertinment, meta, gadget-pr0n, Vermont bloggers (both tech and not), new-music blogs... Not all of them change daily, but many do, so I usually spend about 1..2 hours a day reading stuff. For instance, while the BBC news feeds are pretty much constantly updating, the Netflix new-releases feed will dump about 100 new items every couple of weeks; I can scan through, add the few I want and be on my way.
RSS is better because it's a change delivery mechanism, rather than a static content delivery mechanism (HTML). With an aggregator, it is far more efficient than doing change detection manually if the goal is to stay "current" with published content. Note that it has little to do with mobile devices, and many feeds are full-content (not just headline) feeds.
Sites like that from TFA stopped being interesting a long time ago; the thing we need now is editors. Digg,/., Boing Boing, &c. work because they are actively filtered. They may have different strategies for editorial prioritization of stories. Planet aggregators (Planet GNOME, Planet Apache,...) are still interesting as well because they're naturally filtered to a single topic. Simple feeds (of feeds)^N aren't interesting.
Those sound new, but I don't recall anyway to customize the report. For example, if I only wanted Auto and Health categories in those reports, I don't remember any options to do that.. the report just appeared.
The report has a options... uh... option. This has existed forever, though for some reason it's hard for some people to find. I haven't quite figured out why.
I never knew which one to delete.. what about the.log files? It seems like those needed to be kept too, especially when sometimes it seems that a new.xac and.log were created and were significately smaller than the last 'big' file. I don't recall such a preference, is that new as well?
I believe it was in 1.8 as well. FWIW, the log file is database-style transaction-log of changes to your datafile, whereas the.xac files are straight backup copies.
One big annoyance for me was setting up a recurring transaction; initially, the columns were labled as Funds In and Funds Out, but as soon as I entered ANYTHING it changed to Debit / Credit. Now, I did take an accounting course in HS, did well and remembered much of it, but for some reason I still get confused on Debit vs. Credit.
I'll see if I can't make that at least consistent, if not more clear...
Serious question: in what ways do you find it deficient?
transaction voiding (its supposedly tehre, but I could never find it)
{Transaction > Void Transaction} menu from the register. This is new-in-2.0.
export to QIF (or any other widely used format so that if you chose to leave GC, you can)
The datafile itself is pretty straightforward XML, but you're right -- this would be nice.
home user reports (i.e., pie graph of spending per category, category spending per month, etc etc).
Reports > Income & Expense > Expense Piechart
Reports > Income & Expense > Income/Expense Chart, Expense Barchart
I don't recall any budgeting features.
New in 2.0.
Also, I'm not sure, but I don't think I could download transactions directly from my credit union (without opening a file which is downloaded in some format). The last one... overright the same file.
If OFX, then OFX Direct Connect is supported through AQBanking. Otherwise, no, there's no easy way to either get gnucash to do the download or invoke it as a download handler from the browser.
It seems as if on every save, I'd have 2 or 3 new files with some timestamp added to the filename itself.. so it was Myaccounts_200501234_20050123420050123420050123420 0501234_200501234.xac. How do I know which ones I need or didn't need?
Any but the datafile itself can be removed. But I agree, the backup system is confusing and annoying. There's a preference to clean them after a time-window.
Oh, and get it running on Windows.
That'd be nice.
I"m sure its great if you're using it for business accounting, but it seemed to be aimed much more at that use than home banking.
Hmm. The business features were actually a late addition. It's pretty much set up for personal accounting primarily. We'd love suggestions on how to tailor it better to that application... well, to both, actually.
Seriously, it's kind of ridiculous that this program hasn't been ported yet. I consider OSS that hasn't been written in a cross-platform manner from the start a little dubious.
You don't write much software, do you?
We internally try to use vanilla C constructs for the widest portability.
We abstract multiple standard library uses for portability.
We spent a good deal of effort intentionally restricting ourselves to the FC3/RHEL4 versions of libraries so that the 2.0 release would be able to compile on distros that haven't had any more recent gnucash to upgrade to in the last years since 1.8.
A developer spent a couple of weeks of his own time figuring out changes necessary to get the core code to compile on Windows . In fact, if you check that page, it's hacked modifications of pkgconfig paths and trying to get guile to behave in windows. In fact, very little is gnucash proper.
We'd love to apply patches to get gnucash to compile on windows, or even compilation build failures. But we're certainly not opposed (passively or actively) to getting GnuCash running on windows.
We don't depend on any specific package management tool.
g-wrap was being found by configure, or it wouldn't have finished.
I wish we could help everyone, but Slackware users seem to take the most resource for the least reward. With package managers, things are regular, predictable... problems encountered apply to a whole class of users. With Slack, every system is different. We just don't have the time. Yes, it might be a "philosophical difference" in what a distribution should be, but it's only "philosophical" when it doesn't cross over into our world. In theory, theory is the same as practice. In practice, it's not.
It's less than that now, but it's still a total clusterfsck to get it to run on any distro that doesn't have *gag* package management. The configure/make is still braindead, so it fails with g-wrap errors on slack 10.2
Um. It's not a clusterfsck... you just need to compile the dependent packages in order.
There's a solution to the tedium of that problem: it's called package management. It was invented in the not too long ago as a way to use computers to do a lot of boring, tedious and specific work, like determining and ordering dependencies. Look into it.
Most of our (gnucash's) time-consuming building support regards one distro... can you guess which one?
To mutilate a famous phrase, "Reports of gnucash's dependency problems are greatly exaggerated".
In any case, it's better than before, but you just can't get away from writing desktop software without using other libs to accomplish things like HTML rendering, printing, graphing, xml parsing, &c.
In any case, here's the dependency list from my gentoo install with optional OFX, HBCI support and quotes-fetching turned on; note that most of the packages are standard (zlib, popt) or just "part of gnome" (gtk, glib, gnome*).
IIRC g-wrap-1.9.6 doesn't compile on Solaris 10, and it's a pre-req for gnucash.
We've taken care to make the code base as portable as possible (and to use -- relatively -- really old dependent libs) so it can compile on as many distros/oses as possible. We'd love to have someone helping get it working on Solaris.
While we're not a Full Fledged GNOME Project, the current front-end is built with more than just GTK... gconf, gnome-print, goffice, &c. Developing this level of desktop app needs the set of services provided by a full desktop environment, and while we try to limit what we use, we do use GNOME libs when available.
Of course, nothing's stopping a full KDE front-end over the same core ui-independent engine, except many person-years of development that's gone into the current one.:/
In my opinion, an important issue with GnuCash is that development is fairly slow due to a good chunk of the codebase being LISP, which limits the universe of developers that can work on it. Unless they've changed that with this new version (which would be great!). So it seems like the pace of development compared to other open source products of its age is fairly slow.
A very small fraction of the codebase is lisp.
Development is slow. More developers and/or corporate sponsorship would help. All ~7 active devs have other full-time gigs.
Slashdot Mirror
What we also really need is for people to understand that not only children play video games.
Yes, but the value in sites like /. is *EDITING*. They should filter shit like this so their s/n ratio is high, and thus their feed is valuable. This will bring subscribers and thus attention and thus ad impressions. At present, slashdot posts 90% crap found on other sites 4-9 days ago, and 7% crap like this and 3% interesting stuff, mostly hard-science stuff I'd find if I was still in school or in a non-CS/Real-Science job.
Go do volunteer basic computer literacy session for your local senior center. Don't try to convert them to linux or get them using Firefox or anything dumb like that. Just ask what their problems are, and how you can help. You will quickly understand how broken and unintuitive computer software is.
JAVA will be removed from the *NASDAQ-100 composite index*, but will continue to trade as normal until the company is actually acquired. This point was even mentioned in the press release, so extra points for getting it so (so!) basically wrong.
(Man, /. just continues to accumulate fail. I wonder when it'll implode.)
Funny, it continues to work really well for me every week as I get incremental ugprades that keep my system functional, secure and quite current. I'm more regularly up to date than my big-bang-distro friends, certainly with respect to security issues. I don't have to devote weekends to "upgrading the server", crossing my fingers to make sure it works. And I have multiple options with respect to handling libraries and development packages, and a clear path to mixing stable and unstable packages, and tweaking them as need be. Oh, if I want to add (or remove) support for a feature (e.g., bluetooth), I can easily rebuild every package that provides support for that feature ... without getting the cruft of it when I don't need it.
Quite a "failure", indeed.
(To be clear: Gentoo does have social and architectural problems. But it does have some very compelling points, and the distro as a whole does keep moving on, quite usefully.)
"it's" = "it is". That's all you have to remember. :p
...and quickly watch it tag nearly everything as misspelled. You really want flyspell-prog-mode, but it only checks strings and comments.
True, there's something about both:
...
:) Also, the time wasted in checking the sites that didn't have any changes becomes too great (depending, of course, on how often Your Favorite Sites change).
a/ wanting a lot of new, varied information
b/ not caring about skimming through a bunch of noise to pick out some signal.
A good aggregator makes the second point simpler, but the *varied* part of the former is interesting. It's not about "news" necessarily. I gave the Netflix new releases example, and I just now added the Project Gutenberg newly-available ebooks feed. Some from the forums for a couple of hobbies. A couple of my feeds are new software updates. Multiple feeds of the recent changes of wiki articles: either wikipedia pages I want to make sure maintain correctness, or the GnuCash wiki where I want to see if *anything* changes. Some are (semi-)personal blogs of friends where I just want to see what changes in their lives. Gentoo Linux News, CERT Advisories, new RFCs,
We can easily see the day when common devices -- refridgerators, washing machines, tivo, &c. will have a feed for status. That gets into the filtering and intelligent aggregation aspect, and into The Future, but it's the direction we're headed.
328 is a lot of feeds, granted. But I think the tipping point of "just open them in a bunch of tabs" vs. "aggregator" is around 30 sites. If only because at that point, the tab widths are too small to see which site is which.
I did that too, about 4 years ago. Then it grew to opening up 30 tabs, and having to a) recall what the previous web-page state was and b) identify if anything changed in a mix of differing, sometimes slow-to-load layouts. It was in competition with my slashboxes, which were quite novel and useful at the time. It was great, but then I discovered RSS.
My current (bloglines) subscription list is 327 feeds. They range far and wide: tech commentary, comics, raw news, politics, beer/homebrewing, trashy entertinment, meta, gadget-pr0n, Vermont bloggers (both tech and not), new-music blogs... Not all of them change daily, but many do, so I usually spend about 1..2 hours a day reading stuff. For instance, while the BBC news feeds are pretty much constantly updating, the Netflix new-releases feed will dump about 100 new items every couple of weeks; I can scan through, add the few I want and be on my way.
RSS is better because it's a change delivery mechanism, rather than a static content delivery mechanism (HTML). With an aggregator, it is far more efficient than doing change detection manually if the goal is to stay "current" with published content. Note that it has little to do with mobile devices, and many feeds are full-content (not just headline) feeds.
Sites like that from TFA stopped being interesting a long time ago; the thing we need now is editors. Digg, /., Boing Boing, &c. work because they are actively filtered. They may have different strategies for editorial prioritization of stories. Planet aggregators (Planet GNOME, Planet Apache, ...) are still interesting as well because they're naturally filtered to a single topic. Simple feeds (of feeds)^N aren't interesting.
The report has a options
I believe it was in 1.8 as well. FWIW, the log file is database-style transaction-log of changes to your datafile, whereas the
I'll see if I can't make that at least consistent, if not more clear...
Serious question: in what ways do you find it deficient?
{Transaction > Void Transaction} menu from the register. This is new-in-2.0.
The datafile itself is pretty straightforward XML, but you're right -- this would be nice.
Reports > Income & Expense > Expense Piechart
Reports > Income & Expense > Income/Expense Chart, Expense Barchart
New in 2.0.
If OFX, then OFX Direct Connect is supported through AQBanking. Otherwise, no, there's no easy way to either get gnucash to do the download or invoke it as a download handler from the browser.
Any but the datafile itself can be removed. But I agree, the backup system is confusing and annoying. There's a preference to clean them after a time-window.
That'd be nice.
Hmm. The business features were actually a late addition. It's pretty much set up for personal accounting primarily. We'd love suggestions on how to tailor it better to that application ... well, to both, actually.
You don't write much software, do you?
We internally try to use vanilla C constructs for the widest portability.
We abstract multiple standard library uses for portability.
We spent a good deal of effort intentionally restricting ourselves to the FC3/RHEL4 versions of libraries so that the 2.0 release would be able to compile on distros that haven't had any more recent gnucash to upgrade to in the last years since 1.8.
A developer spent a couple of weeks of his own time figuring out changes necessary to get the core code to compile on Windows . In fact, if you check that page, it's hacked modifications of pkgconfig paths and trying to get guile to behave in windows. In fact, very little is gnucash proper.
We'd love to apply patches to get gnucash to compile on windows, or even compilation build failures. But we're certainly not opposed (passively or actively) to getting GnuCash running on windows.
We don't depend on any specific package management tool.
... problems encountered apply to a whole class of users. With Slack, every system is different. We just don't have the time. Yes, it might be a "philosophical difference" in what a distribution should be, but it's only "philosophical" when it doesn't cross over into our world. In theory, theory is the same as practice. In practice, it's not.
g-wrap was being found by configure, or it wouldn't have finished.
I wish we could help everyone, but Slackware users seem to take the most resource for the least reward. With package managers, things are regular, predictable
Hmm. Should compile well on CentOS 4.3 ... we'd love to know the error...
Um. It's not a clusterfsck
There's a solution to the tedium of that problem: it's called package management. It was invented in the not too long ago as a way to use computers to do a lot of boring, tedious and specific work, like determining and ordering dependencies. Look into it.
Most of our (gnucash's) time-consuming building support regards one distro
To mutilate a famous phrase, "Reports of gnucash's dependency problems are greatly exaggerated".
In any case, it's better than before, but you just can't get away from writing desktop software without using other libs to accomplish things like HTML rendering, printing, graphing, xml parsing, &c.
In any case, here's the dependency list from my gentoo install with optional OFX, HBCI support and quotes-fetching turned on; note that most of the packages are standard (zlib, popt) or just "part of gnome" (gtk, glib, gnome*).
=app-text/docbook-xml-dtd-4.1.2* app-text/docbook-xml-dtd-4.1.2-r6
>=app-text/scrollkeeper-0.3 app-text/scrollkeeper-0.3.14-r2
>=dev-libs/g-wrap-1.3.4 dev-libs/g-wrap-1.9.6
>=dev-libs/glib-2.4.0 dev-libs/glib-2.8.6
>=dev-libs/libxml2-2.5.10 dev-libs/libxml2-2.6.23
>=dev-libs/popt-1.5 dev-libs/popt-1.7-r1
>=dev-libs/slib-2.3.8 dev-libs/slib-2.4.3
>=dev-util/guile-1.6 dev-util/guile-1.6.7
>=gnome-base/gconf-2 gnome-base/gconf-2.12.1
>=gnome-base/libglade-2.4 gnome-base/libglade-2.5.1
>=gnome-base/libgnomeprint-2.10 gnome-base/libgnomeprint-2.12.1
>=gnome-base/libgnomeprintui-2.10 gnome-base/libgnomeprintui-2.12.1
>=gnome-base/libgnomeui-2.4 gnome-base/libgnomeui-2.12.0
>=gnome-extra/gtkhtml-3.6 gnome-extra/gtkhtml-3.8.2
>=sys-apps/sed-4 sys-apps/sed-4.1.4-r1
>=sys-libs/zlib-1.1.4 sys-libs/zlib-1.2.3
>=x11-libs/goffice-0.0.4 x11-libs/goffice-0.1.0
>=x11-libs/gtk+-2.4 x11-libs/gtk+-2.8.12
dev-util/pkgconfig dev-util/pkgconfig-0.20
nls?
sys-devel/gettext sys-devel/gettext-0.14.4
doc?
app-doc/doxygen app-doc/doxygen-1.4.4
app-text/docbook-xsl-stylesheets app-text/docbook-xsl-stylesheets-1.68.1-r1
ofx?
>=dev-libs/libofx-0.7.0 dev-libs/libofx-0.8.0-r1
quotes?
dev-perl/Crypt-SSLeay dev-perl/Crypt-SSLeay-0.51
dev-perl/Finance-Quote dev-perl/Finance-Quote-1.11
dev-perl/HTML-TableExtract dev-perl/HTML-TableExtract-2.06
hbci?
net-libs/aqbanking net-libs/aqbanking-1.6.0_beta
GnuCash has been able to do multiple-split transactions for a long long time.
Exported HTML reports are pretty universal. A good Transaction Report might be sufficient for your accountant.
Not cracked-out at all... works like a charm. I develop gnucash that way sometimes from a windows (and OS X) laptop.
IIRC g-wrap-1.9.6 doesn't compile on Solaris 10, and it's a pre-req for gnucash.
We've taken care to make the code base as portable as possible (and to use -- relatively -- really old dependent libs) so it can compile on as many distros/oses as possible. We'd love to have someone helping get it working on Solaris.
While we're not a Full Fledged GNOME Project, the current front-end is built with more than just GTK ... gconf, gnome-print, goffice, &c. Developing this level of desktop app needs the set of services provided by a full desktop environment, and while we try to limit what we use, we do use GNOME libs when available.
:/
Of course, nothing's stopping a full KDE front-end over the same core ui-independent engine, except many person-years of development that's gone into the current one.
A very small fraction of the codebase is lisp.
Development is slow. More developers and/or corporate sponsorship would help. All ~7 active devs have other full-time gigs.
Budgets: yes (new in 2.0)
Scheduled/recurring transactions: yes (since 1.8)
No, it doesn't have the polish of Commercial Software, but it's improving.
It is available in Fink, though it's not "native" (a-la NeoOffice for OpenOffice).
Only a small fraction of GnuCash is written in guile (scheme). Probably less than 20%.
:)
TIA for some good bugs filed.