By Joe Celko
This can be a kind of books which are maybe great to have as well as whatever higher. Btw, don't overestimate the "for smarties" half within the name: the booklet isn't all that complex: it really is extra like an in depth cookbook with loads of own opinion thrown in (not consistently constant; for instance, in a single position he inveighs opposed to the evils of utilizing sequential-number sequences as fundamental keys -- 'cause a desk isn't a series, you spot, we are conversing approximately units right here, who, by means of definition, are unordered -- okay, nice. Ten pages later he blasts the GUID sort -- why? simply because it isn't inherently sequential and it is tough to identify the gaps within the sequences. yet whats up, why will we care approximately gaps? All we care approximately is that the sphere values be particular, which they're, gaps or no gaps. sounds like GUIDs could be excellent from the set-theoretical standpoint, yet no, he does not like them -- accurately as a result presence of these points, the shortcoming of which he bemoaned one bankruptcy again within the identification variety. it truly is like he wrote those chapters one ten years after the opposite, and forgot what he used to be speaking approximately in a single while writing the other.
The content material (or relatively the reason in the back of it) is excellent. there is a logical development from the overall-schema issues, to tables, and so forth, together with such esoterica as hierarchies and graphs (which is sweet not just, or perhaps now not loads due to the themes themselves, yet simply because great contemporary SQL beneficial properties like CTEs are used much within the pattern code therefore demonstrating their non-trivial use).
The downsides: the most flaw in Celko's writing is that no matter what he writes reads like a two-page magazine article, through which I suggest it is all a perfunctorily dashed-off choice of tidbits; the general constitution is particularly tenuous (for instance, he starts off the hierarchies' bankruptcy by way of announcing they are a type of graphs -- however the graphs right bankruptcy comes afterwards: would it not make extra feel to modify their order within the booklet then?) a few chapters are borrowed from different writers: for instance, the bankruptcy on temporal databases is taken, or fairly squeezed out of Snodgrass's publication (which I occurred to be studying in parallel and hence was once capable of observe that). i am not hinting at plagiarism right here: i am convinced Snodgrass was once conscious of this borrowing and had OK'ed it, yet first it'd be great to say the actual fact of borrowing (I think), and moment, and most vital, if you compress a publication right into a bankruptcy, you gotta do it very rigorously that allows you to preserve the cloth attached, coherent, and transparent. this isn't the case the following (go for the unique: it really is reliable, and will be downloaded at no cost; google at the name).
In common, Celko's writing, whereas now not abhorrent, is generally (though now not all over the place) very sloppy; every little thing reads like a primary draft by no means touched back by way of both the writer or an editor. a good volume of typos, and many doubtful, careless pages that make you fight for that means (not continuously successfully). Mangled French back: Joseph, if there is an accessory over the past 'e' in a masculine kind of a participle, it really is gotta be an accessory aigu, now not accessory grave (feminine easily provides an additional 'e' w/o altering something else); therefore it truly is 'née' now not 'nèe'. might be it is a typo, yet it is continuously repeated thoughout the e-book (and truly found in his different books!). whereas we are right here: "Borland (née Inprise)" is absolutely the wrong way round: Inprise (née Borland).
There's absolute confidence that the writer is a brilliant man, yet writing is not his strong point and he could be much less informal approximately it -- after which, he must also insist that his writer supply reliable editorial oversight. i believe MK is an effective writer, and that i imagine they should have the ability to do a greater task assisting their writers in attaining readability.
Bottom line: i do not remorse having this publication; it really is pleasant and chatty (in a great sense); the inherently dry fabric is livened up a section by way of a sprinkling of curious trivialities; it has been slightly enlightening at the first learn, and time and again invaluable as a reference afterwards. Otoh, it truly is written sloppily and that i consider that, inasmuch as I benefitted from it, I did purely simply because I ensue to have sufficient starting place to atone for its flaws by myself. yet i am not searching for this sort of attempt while analyzing technical books notwithstanding, so 3 stars. may be extra, yet for that the booklet should be aggressively edited and restructured. it is the 3rd version, btw: it'll look that there is been lots of time to just do that.
All in all, i like to recommend it, yet basically half-heartedly: so long as the possible reader knows that this isn't a very complex publication, besides that the analyzing will not be effortless. The book's okay, yet no longer on par with what you get from writers like grey or Date.