Search Results: "ag"

5 December 2008

Joachim Breitner: Good LaTeX Font Overview

I just discovered the nice Survey of Free Math Fonts for TeX and LaTeX (PDF-Version) by Stephen G. Hartke a good overview of the common font packages, with examples, useful for those who have seen enough Computer Modern by now.What I m still missing is a good style guide: When should I use Palatino, when is Utopia a good choice? What are good reasons to use a sans-serif font for the text, and when is that not a good idea? They all look nice to me, but I doubt that the professional typographist leave it all to personal preference.

4 December 2008

Jeff Bailey: Coming up to the third night in the new place...

I found the camera this morning, but I haven't taken pictures of the new place yet. I think that'll have to wait until this weekend.

I'm sitting in the corner of what will be the office space. Right now, it's got the cordless phone, the wifi router, the cable modem, and the Vonage (codec? modem? codec seems likely, how wonderfully archaic). Off to the side is a sewing table of some sort that the former owners left behind which has the power brick for our cell phones and Angie's laptop on it.

We're running laundry for the first time in the washing machine downstairs, and seasoning the new cast iron frying pan. Leif had a bath earlier and actually went down at about the normal time. He woke up at 05h30 this morning, probably due to the lack of curtains. It apparently starts to get bright early, and none of us are sleeping well with three of us in a double.

Our furniture is due to arrive somewhere between the 15th and the 18th. I phoned today and the moving company told us that the stuff hasn't left California yet. Ah well. We've purchased a toaster over and electric kettle so far. We bought the stove, washer and dryer from the former owner. There was a spare fridge downstairs that we're using, but mostly we're using the step outside the kitchen door. We're back in Quebec and it's cold enough to keep things out there. =)

This place is easily the largest place I've lived in since my parents split up when I was 8. Angie and I are really trying hard not to fill it with crap, but actually think through what we need.

I signed Angie and I up for gym memberships today after work. Yesterday was cell phones. Angie picked up some new pots and pans. Tomorrow will be towels and some dishes.

Anyone know if it's better to buy a fridge and dishwasher before Christmas or after? =)

The new home number, cell numbers and such will be in Facebook and Orkut in a moment.

Raphael Geissert: php now has name\spaces support


Finally, the long-awaited release (just alpha atm) of PHP with namespaces support is out, go, get it, and enjoy the sub\spaces\separator before you get\mad\with\it.

Pablo Lorenzzoni: Debian crashing the Internet!?

I just read on Slashdot a pointer to an article about GitTorrent and it made me wonder about these meta-distributed systems. Although I don t share the article s author enthusiasm, it s a really good idea. Anyways, browsing that article, I ended up on another one about Debtorrent. This I ve already read about and decided to keep an eye on for some time now It happens that one line on that article called my attention (and I quote):
At the last major upgrade of Debian/Stable, all the routers at the major International fibreoptic backbone sites across the world redlined for a week. (emphasis added)
Well that was a surprise. Of course I know of the size and importance of Debian, but I thought our systems were more efficient (or that our release-generated-traffic were not of that magnitude). I began googling for a pointer on that Anything: a quote, an URL, a list message anything that could make that claim verifiable: Guess what: I found none! Meanwhile, on Slashdot, other people started following the same subject... So, is it true? Does anyone have any pointers to that? It seems quite unlikely to me But hey! That s just me: maybe the Internet is not that big :-)

Norbert Tretkowski: Dropping MySQL Cluster support from Debian (for a while)

In the next upload of MySQL 5.1.30 to the experimental branch of Debian we are going to drop MySQL Cluster support from the standard MySQL packages. I already wrote something about the future of MySQL Cluster in Debian a while ago, this is just the first step towards a separate mysql-cluster package in Debian.

This change does not (and will never) affect the MySQL packages in etch and lenny, and unstable not until we decide to move MySQL 5.1 from experimental to unstable.

Florian Maier: HP Compaq 2510p linux kernel freeze

I've recently noticed that one of our lab laptops running "testing" and "unstable" <strong>freezes</strong> the system whenever i close
the lid. Additionally, X does not always come up properly and rebooting is (besides typing blindly) the only way to get a
graphical login. The graphics chipset is an intel 955. After a bit of search we found a kernel bugreport, which provides a clue
where the problem occurs. root@orion:~# cat /proc/acpi/video/*/DOS
DOS setting: <0>
To get it working for the HP Compaq 2510p, all you need to do is to set a value of "1" root@orion:~# echo '1' > /proc/acpi/video/*/DOS Then cross your fingers and close the lid. Hopefully your system did no freeze and is still up ;-) At work, we're using a modprobe post-install-Script to work around this which sets the working ACPI _DOS value if a ''ACPIvideo_DOS'' variable is set. This will hopefully be fixed soon in a newer kernel version. Afaik from Debconf8, Bdale also owns one of those notebooks. PS.: *rofl*

Michael Banck: 4 Dec 2008

Opensync updates Some time ago, Opensync-0.38 got released, and it is now available in experimental. The evolution-data-server and the Opie plugins are now available again, as well as the new tomboy (in NEW) and a rewritten google-calendar/contacts plugin. The google plugin requires the new libgcal, which I have just uploaded to NEW. Unfortunately, kitchensync is still not ported to latest Opensync-0.3x (and got dropped for KDE4.2), so one still needs to use the command-line msynctool program. Also not ported are the (KDE3) kdepim and the currently under development Akonadi plugin. Other important plugins missing for 0.38 are the Windows Mobile, Blackberry, Palm and IRMC plugins. I tried to suggest making 0.38.x point releases including more ported plugins, but it seems development is turning towards 0.39 already, and yet some more API changes were done, this time mostly removing unnecessary interfaces, which should be a good thing in the long term. Some other good news is that there are now weekly IRC meetings of the Opensync developers, so there should be steadier progress towards Opensync-0.40 from now on. Unfortunately, I was mostly absent during all of the three meetings so far. Along with Opensync-0.38, libsyncml saw a new major release 0.5.0 which should fix lots of bugs and provide better support for mobiles. However, lots of problems with syncml were due to bugs in the wbxml2 library. Michael Bell has hopefully found the most critical ones and I have uploaded a new wbxml2-0.9.2 to unstable today which I hope will get into lenny soon. The main problem with wbxml2 over the last year was a unresponsive/MIA upstream; however, recently wbxml2 maintainership got tranferred to the opensync project and moved to its Trac. Michael Bell has been fixing most of the outstanding issues and is currently preparing a 0.10.0 release, so this project should be back on track now.

Gintautas Miliauskas: Python 3.0

Python 3.0 (also known as Python 3000) has been released yesterday, so I thought I'd give it a try. It turns out that a recent release candidate is in official Ubuntu 8.10 (intrepid) repositories, so I did not even need to compile anything: apt-get install python3.0 and we're done. You may want to read the summary of changes from Python 2 beforehand. Here's a quick demo of unicode identifiers:
>>>   = 2
>>>   =   +  
>>>  
4
>>> import math
>>>   = math.pi
>>>   =  /3
>>> math.sin( )
0.8660254037844386
This isn't quite APL yet, but we are getting closer!

David Welton: The Economics of Programming Languages - in the 1950's

I was having a look at this paper on the development of Fortran: http://www.softwarepreservation.org/projects/FORTRAN/paper/p165-backus.pdf (which I found here ) And I noticed the following quote:
1.2. The Economics of Programming Another factor which influenced the development of FORTRAN was the economics of programming in 1954. The cost of programmers associated with a computer center was usually at least as great as the cost of the computer itself. (This fact follows from the aver- age salary-plus-overhead and number of programmers at each center and from the computer rental figures.) In addition, from one-quarter to one-half of the computer's time was spent in debugging. Thus programming and debugging accounted for as much as three-quarters of the cost of operating a computer; and obviously, as computers got cheaper, this situation would get worse. This economic factor was one of the prime motivations which led me to propose the FORTRAN project in a letter to my boss, Cuthbert Hurd, in late 1953 (the exact date is not known but other facts suggest December 1953 as a likely date). I believe that the economic need for a system like FORTRAN was one reason why IBM and my successive bosses, Hurd, Charles DeCarlo, and John McPherson, provided for our constantly expanding needs over the next five years without ever asking us to project or justify those needs in a formal budget.
As Hal Varian says, "Ignore basic economic principles at your own risk. Technology changes. Economic laws do not." I would continue to bet on programming languages that save programmers time, or otherwise make them more productive.

David Welton: The Economics of Programming Languages - in the 1950's

I was having a look at this paper on the development of Fortran: http://www.softwarepreservation.org/projects/FORTRAN/paper/p165-backus.pdf (which I found here ) And I noticed the following quote:
1.2. The Economics of Programming Another factor which influenced the development of FORTRAN was the economics of programming in 1954. The cost of programmers associated with a computer center was usually at least as great as the cost of the computer itself. (This fact follows from the aver- age salary-plus-overhead and number of programmers at each center and from the computer rental figures.) In addition, from one-quarter to one-half of the computer's time was spent in debugging. Thus programming and debugging accounted for as much as three-quarters of the cost of operating a computer; and obviously, as computers got cheaper, this situation would get worse. This economic factor was one of the prime motivations which led me to propose the FORTRAN project in a letter to my boss, Cuthbert Hurd, in late 1953 (the exact date is not known but other facts suggest December 1953 as a likely date). I believe that the economic need for a system like FORTRAN was one reason why IBM and my successive bosses, Hurd, Charles DeCarlo, and John McPherson, provided for our constantly expanding needs over the next five years without ever asking us to project or justify those needs in a formal budget.
As Hal Varian says, "Ignore basic economic principles at your own risk. Technology changes. Economic laws do not." I would continue to bet on programming languages that save programmers time, or otherwise make them more productive.

Jon Dowland: nautilus sums

I've got a ISO image that I wish to burn to DVD. I've got a copy burned already which has the same MD5 sum as the ISO file. I've just attempted to burn it another two times, both using nautilus cd burner (well, right click and 'burn image'). In both cases, the sum ended up being different, but strangely, the sum was the same for both of these subsequent burns. Has anyone else experienced problems getting 1:1 burns with nautilus cd burner?

Jon Dowland: planet spamming

I've just migrated hosts for my website and managed to spam at least one aggregator in the process. (I think this is because git does not preserve timestamp metadata for blobs but I haven't investigated properly yet). Apologies to those affected. Luckily, I have my feed configured to only show the most recent 5 items, so the impact was quite small. I think this should be the default for applications that generate feeds for precicely this reason: It is for blosxom (#373570) and pyblosxom (#373658), and ikiwiki specifies a limit, at least. Asides from feeds, I think blog web pages that display an enormous number of posts on the front page (or inside category pages) are demonstrating some pretty bad usability. I thought that Jakob Nielsen may have included this in his Weblog Usability: The Top Ten Design Mistakes, but he didn't. Thomas Watson writes we have celebrated with finally scrapping our blog archive. Instead we are now just showing the latest 50 articles on the front page, followed by a link to the previous 50., which pretty much the opposite of what I'd advise. The point he makes, though, is that the alternative (an ever-expanding "archive", accessed by calendar) is bad form, too. That is one of the points in Nielsen's list. Jonathan Boutelle describes an interesting compromise which he calls the "Mullet style" - a short list of full posts followed by a list of older posts, containing only their titles and date.

David Welton: Won't fix... why?

I discovered a minor Rails annoyance today, and, looking around for an answer to the problem, I found this, which is a perfect description of the issue, and includes a fix: http://dev.rubyonrails.org/ticket/5123 Unfortunately it has marked as "wont fix" with no reason specified. Perhaps there is a reasonable answer that was perhaps discussed elsewhere, and just didn't make it into the bug report, making it invisible to people like me who find it via Google, which is pretty annoying. The other explanation is more along the lines of "don't wanna" (or fuck you, as the case may be), but at least come out and say so. "We're not going to do it, as it's not part of our English-centric vision" or what have you. Being a native English speaker, even when doing stuff for Italian or Austrian or whatever clients, I tend to code in English, because it's 1) fast, 2) easy and 3) it's pretty much universally understood by programmers. However, I can see the case for specific things to be in a particular language, and on the surface, this one line patch makes that a little bit easier, and doesn't appear to be costly, so why not? Just closing the bugs of people who have taken the time to try and help (this is clearly not just a "dude, it's, uh, broke" bug) is the twin of the equally annoying people who don't bother to take 30 seconds to send an email or bug report about software they're using, and both are things I greatly dislike.

David Welton: Won't fix... why?

I discovered a minor Rails annoyance today, and, looking around for an answer to the problem, I found this, which is a perfect description of the issue, and includes a fix: http://dev.rubyonrails.org/ticket/5123 Unfortunately it has marked as "wont fix" with no reason specified. Perhaps there is a reasonable answer that was perhaps discussed elsewhere, and just didn't make it into the bug report, making it invisible to people like me who find it via Google, which is pretty annoying. The other explanation is more along the lines of "don't wanna" (or fuck you, as the case may be), but at least come out and say so. "We're not going to do it, as it's not part of our English-centric vision" or what have you. Being a native English speaker, even when doing stuff for Italian or Austrian or whatever clients, I tend to code in English, because it's 1) fast, 2) easy and 3) it's pretty much universally understood by programmers. However, I can see the case for specific things to be in a particular language, and on the surface, this one line patch makes that a little bit easier, and doesn't appear to be costly, so why not? Just closing the bugs of people who have taken the time to try and help (this is clearly not just a "dude, it's, uh, broke" bug) is the twin of the equally annoying people who don't bother to take 30 seconds to send an email or bug report about software they're using, and both are things I greatly dislike.

MJ Ray: Do Your Shop Photos Leak?

If you’ve set up an online shop, are your product photos saying more about you than you intended?
“there’s a huge amount of potentially privacy-sensitive metadata in your typical JPEG as generated by your camera (including camera type, settings, date/time, maybe even GPS coordinates of your location, etc).” Source: jhead - List and modify EXIF fields in JPEG photos [Uwe Hermann's blog]
Even worse, if you’ve been using images from your suppliers or the manufacturers without permission, the metadata can be a dead giveaway. You really shouldn’t be using them without permission anyway, though. If you’re using your own photos, want to remove the camera data and you’re not comfortable using the command line, most graphics editors like GIMP allow you to edit or remove that extra information too (I think it’s on Save As… OK… Advanced Options and untick “Save EXIF data” in GIMP). It’s just a bit slower to convert lots of images with a GUI. (I’d usually do that sort of thing from the command line for TTLLP customers.) One fringe benefit is that removing the thumbnails and EXIF makes the images a bit smaller, so they’re quicker to download!

3 December 2008

Romain Francoise: Cleaner

In the productive procrastination department: I repackaged rcs using debhelper (it feels like the future) and while I was at it, I implemented MadCoder's git maintenance scheme where changes from upstream are maintained in a rebased integration branch and serialized as patches in the Debian branch (master in my case) using git format-patch. It's pretty nice:Of course this is a very low-maintenance package (the last upstream release was 13 years ago) so it doesn't matter very much which maintenance strategy it uses, but after implementing it I'm pretty confident that it works and I might switch my other packages to it.

If you're interested in knowing more about clever ways to use modern VCS for packaging, you may want to join the vcs-pkg mailing-list.

Obey Arthur Liu: FOSDEM 2009

I'm going to FOSDEM, the Free and Open Source Software Developers' European Meeting Are you ? I was also entertaining the idea of giving a lightning talk in the Debian room (not as part of the general lightning talks) about aptitude-gtk. I certainly don’t feel like talking for a whole hour talk or what. I don’t have that much interesting stuff to say. Maybe we could do it with other 2008 european GSoCers. That could be useful as we didn’t do much publicity-wise (see) in a timely maneer. Certainly I should write a mail about that to the soc-coordination list about this or ramble about it on the planet. :) Oh, I’ll get around it. So any Debian GSoCers going to FOSDEM ?

Christian Perrier: Why not so much incoming mail today?

Waking up this morning, I noticed a low traffic in my incoming mail in "Debian work" mailboxes (PTS mailboxes, personal email, bugs, -release, etc.). Then I read -devel and understood. Everything is there. We have a new flamewar (apparently about NEW processing). It's interesting to see that the closer we get from a release, the biggest flamewars we are doing. Sounds like everybody is boring (have you notived the strong drop in the bug report rate?). Crap, there's a lot of work left to do, to release the damn Lenny: For $ DEITY 's sake, could we do that, so that developers stop boring, then go back to their development duties and the number of flamewars lowers down again (maybe after the squeeze release goals discussion which will probably be "interesting", also)? Please...your home page, these days, should really be http://bts.turmzimmer.net...

Matthew Garrett: Advantage of Linux over Solaris

In response to this:

Idle power draw of Fedora 10: 100W
Idle power draw of Opensolaris 2008-11: 135W

Sure, there are other arguments. But with energy prices the way they are, choosing Solaris is choosing to spend significantly more money on your hosting. Remember to factor that into your TCO calculations.

(Testing done on a 4-way AMD with a Radeon X1900 graphics card, which is what I happen to have to hand. Out of the box configuration.)

David Bremner: Using GLPK from C++

Recently I suggested to some students that they could use the Gnu Linear Programming Toolkit from C++. Shortly afterwards I thought I had better verify that I had not just sent people on a hopeless mission. To test things out, I decided to try using GLPK as part of an ongoing project with Lars Schewe The basic idea of this example is to use glpk to solve an integer program with row generation. The main hurdle (assuming you want to actually write object oriented c++) is how to make the glpk callback work in an object oriented way. Luckily glpk provides a pointer "info" that can be passed to the solver, and which is passed back to the callback routine. This can be used to keep track of what object is involved. Here is the class header
#ifndef GLPSOL_HH
#define GLPSOL_HH
#include "LP.hh"
#include "Vektor.hh"
#include "glpk.h"
#include "combinat.hh"
namespace mpc  
  class  GLPSol : public LP  
  private:
    glp_iocp parm;
    static Vektor<double> get_primal_sol(glp_prob *prob);
    static void callback(glp_tree *tree, void *info);
    static int output_handler(void *info, const char *s);
  protected:
    glp_prob *root;
  public:
    GLPSol(int columns);
    ~GLPSol()  ;
    virtual void rowgen(const Vektor<double> &candidate)  ;
    bool solve();
    bool add(const LinearConstraint &cnst);
   ;
 
#endif
The class LP is just an abstract base class (like an interface for java-heads) defining the add method. The method rowgen is virtual because it is intended to be overridden by a subclass if row generation is actually required. By default it does nothing. Notice that the callback method here is static; that means it is essentially a C function with a funny name. This will be the function that glpk calls when it wants from help.
#include <assert.h>
#include "GLPSol.hh"
#include "debug.hh"
namespace mpc 
  GLPSol::GLPSol(int columns)  
    // redirect logging to my handler
    glp_term_hook(output_handler,NULL);
    // make an LP problem
    root=glp_create_prob();
    glp_add_cols(root,columns);
    // all of my variables are binary, my objective function is always the same
    //  your milage may vary
    for (int j=1; j<=columns; j++) 
      glp_set_obj_coef(root,j,1.0);
      glp_set_col_kind(root,j,GLP_BV);
     
    glp_init_iocp(&parm);
    // here is the interesting bit; we pass the address of the current object
    // into glpk along with the callback function
    parm.cb_func=GLPSol::callback;
    parm.cb_info=this;
   
  int GLPSol::output_handler(void *info, const char *s) 
    DEBUG(1) << s;
    return 1;
   
  Vektor<double> GLPSol::get_primal_sol(glp_prob *prob) 
    Vektor<double> sol;
    assert(prob);
    for (int i=1; i<=glp_get_num_cols(prob); i++) 
      sol[i]=glp_get_col_prim(prob,i);
     
    return sol;
   
  // the callback function just figures out what object called glpk and forwards
  // the call. I happen to decode the solution into a more convenient form, but 
  // you can do what you like
  void GLPSol::callback(glp_tree *tree, void *info) 
    GLPSol *obj=(GLPSol *)info;
    assert(obj);
    switch(glp_ios_reason(tree)) 
    case GLP_IROWGEN:
      obj->rowgen(get_primal_sol(glp_ios_get_prob(tree)));
      break;
    default:
      break;
     
   
  bool GLPSol::solve(void)    
    int ret=glp_simplex(root,NULL);
    if (ret==0) 
      ret=glp_intopt(root,&parm);
    if (ret==0)
      return (glp_mip_status(root)==GLP_OPT);
    else
      return false;
   
  bool GLPSol::add(const LinearConstraint&cnst) 
    int next_row=glp_add_rows(root,1);
    // for mysterious reasons, glpk wants to index from 1
    int indices[cnst.size()+1];
    double coeff[cnst.size()+1];
    DEBUG(3) << "adding " << cnst << std::endl;
    int j=1;
    for (LinearConstraint::const_iterator p=cnst.begin();
         p!=cnst.end(); p++) 
      indices[j]=p->first;
      coeff[j]=(double)p->second;
      j++;
     
    int gtype=0;
    switch(cnst.type()) 
    case LIN_LEQ:
      gtype=GLP_UP;
      break;
    case LIN_GEQ:
      gtype=GLP_LO;
      break;
    default:
      gtype=GLP_FX;
     
    glp_set_row_bnds(root,next_row,gtype,       
                       (double)cnst.rhs(),(double)cnst.rhs());
    glp_set_mat_row(root,
                    next_row,
                    cnst.size(),
                    indices,
                    coeff);
    return true;
   
 
All this is a big waste of effort unless we actually do some row generation. I'm not especially proud of the crude rounding I do here, but it shows how to do it, and it does, eventually solve problems.
#include "OMGLPSol.hh"
#include "DualGraph.hh"
#include "CutIterator.hh"
#include "IntSet.hh"
namespace mpc 
  void OMGLPSol::rowgen(const Vektor<double>&candidate) 
    if (diameter<=0) 
      DEBUG(1) << "no path constraints to generate" << std::endl;
      return;
     
    DEBUG(3) << "Generating paths for " << candidate << std::endl;
  // this looks like a crude hack, which it is, but motivated by the
  // following: the boundary complex is determined only by the signs
  // of the bases, which we here represent as 0 for - and 1 for +
    Chirotope chi(*this);
    for (Vektor<double>::const_iterator p=candidate.begin();
         p!=candidate.end(); p++) 
      if (p->second > 0.5)  
        chi[p->first]=SIGN_POS;
        else  
        chi[p->first]=SIGN_NEG;
       
     
    BoundaryComplex bc(chi);
    DEBUG(3) << chi;
    DualGraph dg(bc);
    CutIterator pathins(*this,candidate);
    int paths_found=
      dg.all_paths(pathins,
                   IntSet::lex_set(elements(),rank()-1,source_facet),
                   IntSet::lex_set(elements(),rank()-1,sink_facet),
                   diameter-1);
    DEBUG(1) << "row generation found " << paths_found << " realized paths\n";
    DEBUG(1) << "effective cuts: " << pathins.effective() << std::endl;
   
  void OMGLPSol::get_solution(Chirotope &chi)  
    int nv=glp_get_num_cols(root);
    for(int i=1;i<=nv;++i)  
      int val=glp_mip_col_val(root,i);
      chi[i]=(val==0 ? SIGN_NEG : SIGN_POS);
     
   
 
So ignore the problem specific way I generate constraints, the key remaining piece of code is CutIterator which filters the generated constraints to make sure they actually cut off the candidate solution. This is crucial, because row generation must not add constraints in the case that it cannot improve the solution, because glpk assumes that if the user is generating cuts, the solver doesn't have to.
#ifndef PATH_CONSTRAINT_ITERATOR_HH
#define PATH_CONSTRAINT_ITERATOR_HH
#include "PathConstraint.hh"
#include "CNF.hh"
namespace mpc  
  class CutIterator : public std::iterator<std::output_iterator_tag,
                                                      void,
                                                      void,
                                                      void,
                                                      void> 
  private:
    LP& _list;
    Vektor<double> _sol;
    std::size_t _pcount;
    std::size_t _ccount;
  public:
    CutIterator (LP& list, const Vektor<double>& sol) : _list(list),_sol(sol), _pcount(0), _ccount(0)  
    CutIterator& operator=(const Path& p)  
      PathConstraint pc(p);
      _ccount+=pc.appendTo(_list,&_sol);
      _pcount++;
      if (_pcount %10000==0)  
        DEBUG(1) << _pcount << " paths generated" << std::endl;
       
      return *this;
     
    CutIterator& operator*()  return *this; 
    CutIterator& operator++()  return *this; 
    CutIterator& operator++(int)  return *this; 
    int effective()   return _ccount;  ;
   ;
 
#endif
Oh heck, another level of detail; the actual filtering actually happens in the appendTo method the PathConstraint class. This is just computing the dot product of two vectors. I would leave it as an exercise to the readier, but remember some fuzz is neccesary to to these kinds of comparisons with floating point numbers. Eventually, the decision is made by the following feasible method of the LinearConstraint class.
 bool feasible(const
Vektor<double> & x)  double sum=0; for (const_iterator
p=begin();p!=end(); p++)  sum+= p->second*x.at(p->first);  
      switch (type()) 
      case LIN_LEQ:
        return (sum <= _rhs+epsilon);
      case LIN_GEQ:
        return (sum >= _rhs-epsilon);
    default:
      return (sum <= _rhs+epsilon) &&
        (sum >= _rhs-epsilon);
       
     

Next.

Previous.